voice_actor_recognition.ipynb

{
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/kimurash/voice_actor_recognition/blob/main/voice_actor_recognition.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "KZzA0ys5LiqW"
      },
      "source": [
        "# 声優100人の**話者認識**"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "UjqPZctyswnx"
      },
      "source": [
        "## 実験の動機"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "OkMU0x22s23J"
      },
      "source": [
        "長年アニメを見続けていると，よく出演されている声優さんを識別できるようになってくる．\n",
        "これは声優さんの声の特徴を脳が無意識に学習するためだと考えられるが，同じことを機械学習でも実現できるはずだと思ったのがこの実験の動機である．"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "25VCYJjmUBGS"
      },
      "source": [
        "## 学習データの用意"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "学習データとして，東京大学の[高道 慎之介 助教](https://sites.google.com/site/shinnosuketakamichi/home)がCC BY-SAライセンスに基づいて公開している [JVS corpus](https://sites.google.com/site/shinnosuketakamichi/research-topics/jvs_corpus) という音声コーパスのうち，[日本声優統計学会](https://voice-statistics.github.io/)が独自に構築した音素バランス文を，100人のプロフェッショナル話者 (声優・俳優など) が読み上げた音声が収録された `parallel100` を採用する．"
      ],
      "metadata": {
        "id": "3YTQ0G-PtoEO"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "Colaboratory に Google Drive をマウントすることで保管してある音声データを読み込めるようにする．"
      ],
      "metadata": {
        "id": "GmYkfNWplSUd"
      }
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "OgJi52l6SONY",
        "outputId": "fa659717-d9db-47b1-cc24-89953d4936ca"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Mounted at /content/drive\n"
          ]
        }
      ],
      "source": [
        "from google.colab import drive\n",
        "drive.mount('/content/drive')\n",
        "\n",
        "# 本実験に用いるデータが保管されたディレクトリへのパス\n",
        "base_path = './drive/MyDrive/Colab Notebooks/personal/voice_actor_recog/'"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "2sxdGRJjTzTU"
      },
      "source": [
        "## 特徴量の抽出"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "fJWJ721V1ftT"
      },
      "source": [
        "今回の実験ではMFCC: Mel Frequrency Cepstrum Coefficientと呼ばれる特徴量を採用した．以降では，MFCCがどういう特徴量であるか説明することで**音声認識**における有効性を示し，手順に沿って実際に計算してみる．"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "0gulpd12tHMh"
      },
      "source": [
        "### **音声認識**に有効な特徴量"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "ubJ42NS_13Dd"
      },
      "source": [
        "**音声認識**を行う場合は，音素を区別するための情報が必要である．人間の発声メカニズムを解き明かすことで，音素の違いは共振周波数の違いとなって現れることが分かる．そこで，音声信号をフーリエ変換によってスペクトルに変換し，スペクトルの山 (フォルマント) を観測することで共振周波数の情報を得るという方法が考えられる．しかし，スペクトルには共振周波数の情報 (スペクトルの概形) だけでなく，基本周波数の情報 (微細な振動) も乗っているため，この情報をうまく取り除く必要がある．このため，スペクトルをさらにフーリエ変換して低周波成分と高周波成分に分離し，低周波成分のみを使用するというのがMFCCの基本的なアイデアである．"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "MZSzkNNMVUmG"
      },
      "source": [
        "ここまでの説明を踏まえて，1フレームの音声信号からMFCCを求める手順を以下に示す．フレームとは，音声信号から一部分を切り出すときの単位であり，フレーム長は通常25ms程度に設定する[1]．\n",
        "\n",
        "1. フレーム両端での急激な変化を避けるために窓関数をかける．\n",
        "2. フレームの音声波形をフーリエ変換する．  \n",
        "3. 人間の聴覚特性を反映させるためにメルフィルタバンクをかける．\n",
        "3. 3.で得られたスペクトルをさらに離散コサイン変換し，低周波成分と高周波成分に分ける．\n",
        "5. 離散コサイン変換の低次の係数から12次元を取り出す．\n",
        "\n",
        "**音声認識**では，フレームを10ms程度ずらしながら上記の処理を行う．これは人間の聴覚の時間分解能に対応している[1]．"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "POOzoMhT1zZ2"
      },
      "source": [
        "### 実験の方針"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "IC9C0OIS28Dg"
      },
      "source": [
        "ここまで長々とMFCCについて説明してきたが，本実験のタスクは**音声認識**ではなく**話者認識**であることを思い出して欲しい．同じ内容を読み上げた音声ファイルから話者を識別するには，「何を話しているか」よりも「どのように話しているか」の方が重要である．スペクトルに乗っている基本周波数の情報は声の高さ(イントネーション)を表しており，この情報は離散フーリエ変換後の高周波成分に現れる．したがって，本実験タスクにおいては，MFCCの低次元よりも高次元の情報を用いた方がうまく識別できると考えられる．\n",
        "\n",
        "本実験では，MFCCの低次から12次元とった場合と，高次から12次元とった場合で識別性能を比較してみようと思う．"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "CZuOyg6nT3Iu"
      },
      "source": [
        "### MFCCを計算してみる"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "本節では [JVS corpus](https://sites.google.com/site/shinnosuketakamichi/research-topics/jvs_corpus) に用意されているサンプルファイル `sample_jvs001.wav` を用いて実際にMFCCを計算してみる．"
      ],
      "metadata": {
        "id": "bKZjUgk5kpyN"
      }
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "dDcPeucq_J1f"
      },
      "source": [
        "必要なライブラリを読み込む．"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "VWpookG61vwX"
      },
      "outputs": [],
      "source": [
        "import numpy as np\n",
        "import matplotlib.pyplot as plt\n",
        "import pandas as pd\n",
        "np.set_printoptions(suppress=True, precision=7)\n",
        "\n",
        "import os\n",
        "import soundfile as sf"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "jDfqMH7B_RdT"
      },
      "source": [
        "#### 1フレームを切り出す"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 297
        },
        "id": "AYlsQfQY271w",
        "outputId": "6024e954-7cfb-4103-be4c-d309c39f2fb2"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "sampling rate:24000\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ],
      "source": [
        "data, fs = sf.read(base_path + 'sample_jvs001.wav')\n",
        "print(f'sampling rate:{fs}')\n",
        "\n",
        "time = np.arange(0, len(data)/fs, 1/fs)\n",
        "\n",
        "center = len(data) // 2\n",
        "cuttime = 0.04\n",
        "x = data[int(center - (cuttime / 2) * fs) : int(center + (cuttime/2) * fs)]\n",
        "t = time[int(center - (cuttime / 2) * fs) : int(center + (cuttime/2) * fs)]\n",
        "\n",
        "plt.plot(t * 1000, x)\n",
        "plt.xlabel(\"time [ms]\")\n",
        "plt.ylabel(\"amplitude\")\n",
        "plt.show()"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "1Ynmgl3Z_Yju"
      },
      "source": [
        "#### 窓関数をかけてフーリエ変換する"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 279
        },
        "id": "VHx-PhKn5RoW",
        "outputId": "cd2f8e2d-57f2-47ba-8169-6e4d9cb325f6"
      },
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ],
      "source": [
        "# ハミング窓をかける\n",
        "hamming = np.hamming(len(x))\n",
        "x = x * hamming\n",
        "\n",
        "N = 2048 # FFTのサンプル数(2の累乗で高速化)\n",
        "\n",
        "# ナイキスト周波数以下のデータを取り出す\n",
        "spectre = np.abs(np.fft.fft(x, N))[:N // 2] # フーリエ変換\n",
        "fscale = np.fft.fftfreq(N, d = 1.0 / fs)[:N // 2] # フーリエ変換の周波数を取得\n",
        "\n",
        "plt.plot(fscale, spectre)\n",
        "plt.xlabel(\"frequency [Hz]\")\n",
        "plt.ylabel(\"amplitude spectrum\")\n",
        "plt.show()"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "53pHD6Cq_i_g"
      },
      "source": [
        "#### メルフィルタバンクをかける"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "メルフィルタバンクを求める手順を以下に示す．\n",
        "\n",
        "1. ナイキスト周波数をメル尺度に変換する．\n",
        "2. メル尺度上で0からナイキスト周波数までの区間をチャンネル数で等分割する．\n",
        "3. 2.で求めた分割を周波数に変換する．\n",
        "4. 分割の各点を底辺の中点とする三角フィルタを作成する．"
      ],
      "metadata": {
        "id": "qTk-CuYUklP8"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "周波数とメル尺度を変換する関数をあらかじめ定義しておく．"
      ],
      "metadata": {
        "id": "0LhOZeC2h9Zk"
      }
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "46PdwrAE_pte"
      },
      "outputs": [],
      "source": [
        "def hz2mel(f):\n",
        "    \"\"\"Hzをmelに変換\"\"\"\n",
        "    return 2595 * np.log(f / 700.0 + 1.0)\n",
        "\n",
        "def mel2hz(m):\n",
        "    \"\"\"melをHzに変換\"\"\"\n",
        "    return 700 * (np.exp(m / 2595) - 1.0)"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "メルフィルタバンクを作成する関数を定義する．\n",
        "\n",
        "関数内で使われている変数をアルゴリズムと共に図示した画像を用意した．  \n",
        "関数の定義と照らし合わせながら読んで，理解の一助として欲しい．"
      ],
      "metadata": {
        "id": "7oMvXlPtlM73"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "from IPython.display import Image,display_png\n",
        "\n",
        "img_path = base_path + '/mel_filter_bank.png'\n",
        "display_png(Image(data=img_path, width=550, height=1000))"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "id": "nNdBNupvYUp0",
        "outputId": "330cf3fb-3c84-4220-bdbb-a85ddd9b0057"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n"
          },
          "metadata": {
            "image/png": {
              "width": 550,
              "height": 1000
            }
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "4TMFOQ1t_ush"
      },
      "outputs": [],
      "source": [
        "def mel_filter_bank(fs, N, n_channel):\n",
        "    \"\"\"メルフィルタバンクを作成\"\"\"\n",
        "    # ナイキスト周波数[Hz]\n",
        "    fmax = fs / 2\n",
        "    # ナイキスト周波数[mel]\n",
        "    melmax = hz2mel(fmax)\n",
        "    # 周波数インデックスの最大値\n",
        "    nmax = N // 2\n",
        "    # 周波数解像度(周波数インデックス1あたりの[Hz]幅)\n",
        "    df = fs / N\n",
        "\n",
        "    # メル尺度における各フィルタの中心周波数を求める\n",
        "    dmel = melmax / (n_channel + 1)\n",
        "    melcenters = np.arange(1, n_channel + 1) * dmel\n",
        "\n",
        "    # 各フィルタの中心周波数を[Hz]に変換\n",
        "    fcenters = mel2hz(melcenters)\n",
        "    # 各フィルタの中心周波数を周波数インデックスに変換\n",
        "    indexcenter = np.round(fcenters / df)\n",
        "\n",
        "    # 各フィルタの開始位置のインデックス\n",
        "    indexstart = np.hstack(([0], indexcenter[0:n_channel - 1]))\n",
        "    # 各フィルタの終了位置のインデックス\n",
        "    indexstop = np.hstack((indexcenter[1:n_channel], [nmax]))\n",
        "\n",
        "    filterbank = np.zeros((n_channel, nmax))\n",
        "    # print(indexstop)\n",
        "\n",
        "    for c in range(0, n_channel): # c: channel\n",
        "        # 三角フィルタの左の直線の傾きから点を求める\n",
        "        increment= 1.0 / (indexcenter[c] - indexstart[c]) # 右上がりの傾き\n",
        "\n",
        "        for i in range(int(indexstart[c]), int(indexcenter[c])):\n",
        "            filterbank[c, i] = (i - indexstart[c]) * increment\n",
        "\n",
        "        # 三角フィルタの右の直線の傾きから点を求める\n",
        "        decrement = 1.0 / (indexstop[c] - indexcenter[c]) # 右下がりの傾き\n",
        "        for i in range(int(indexcenter[c]), int(indexstop[c])):\n",
        "            filterbank[c, i] = 1.0 - ((i - indexcenter[c]) * decrement)\n",
        "\n",
        "    return filterbank, fcenters"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "上で定義した関数を用いてメルフィルタバンクを作成し，描画する．"
      ],
      "metadata": {
        "id": "8LxcENARlVbD"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "n_channel = 20 # メルフィルタバンクのチャネル数\n",
        "df = fs / N    # 周波数解像度(周波数インデックス1あたりの[Hz]幅)\n",
        "filterbank, fcenters = mel_filter_bank(fs, N, n_channel)\n",
        "\n",
        "# メルフィルタバンクのプロット\n",
        "for c in np.arange(0, n_channel):\n",
        "    plt.plot(np.arange(0, N // 2) * df, filterbank[c])\n",
        "\n",
        "plt.title('Mel Filter Bank')\n",
        "plt.xlabel('Frequency[Hz]')\n",
        "plt.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 295
        },
        "id": "sHCoTYaVlLln",
        "outputId": "ba87e4d2-bf19-4309-9484-a757cddca826"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "xSEcGe-7JeNN"
      },
      "outputs": [],
      "source": [
        "# 振幅スペクトルにメルフィルタバンクを適用\n",
        "mspectre = np.dot(spectre, filterbank.T)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 334
        },
        "id": "lfvGhnq3JjJy",
        "outputId": "2e0fb64c-82e4-436a-e138-dd388a7d9c69"
      },
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 936x360 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ],
      "source": [
        "# 元の振幅スペクトルとフィルタバンクをかけて圧縮したスペクトルを表示\n",
        "plt.figure(figsize=(13, 5))\n",
        "\n",
        "plt.plot(fscale, 10 * np.log10(spectre), label='Original Spectrum')\n",
        "plt.plot(fcenters, 10 * np.log10(mspectre), \"o-\", label='Mel Spectrum')\n",
        "plt.xlabel(\"frequency[Hz]\")\n",
        "plt.ylabel('Amplitude[dB]')\n",
        "plt.legend()\n",
        "plt.show()"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "CUVaJ9-0K3n1"
      },
      "source": [
        "#### 離散コサイン変換する"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "e0BaLjJCK9_L",
        "outputId": "f30654b2-32f7-44d4-919a-ea7390a44710"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "(20,)\n",
            "[13.3190939 24.2771886  6.9022144 -3.5544651 12.9122567  0.5574299\n",
            " -1.8089861  1.496971  -6.083573   0.5448475 -2.8065642 -0.0265315\n",
            "  0.8130684  0.1490588 -0.9149896 -1.70858    2.6811971 -0.3079438\n",
            "  1.3090417 -1.841087 ]\n"
          ]
        }
      ],
      "source": [
        "from scipy.fftpack.realtransforms import dct\n",
        "\n",
        "cepstre = dct(10 * np.log10(mspectre), type=2, norm=\"ortho\", axis=-1)\n",
        "print(cepstre.shape)\n",
        "print(cepstre)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "NDrRbftxUnbP"
      },
      "source": [
        "### Librosaを用いたMFCCの算出"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "前節では音声データの1フレームに対してMFCCを求めたが，[Librosa](https://librosa.org/doc/latest/index.html) というライブラリを利用すると，各フレームのMFCCを簡単に計算できる．以下のセルでは，全音声データに対して，各フレームのMFCCを Librosa の `librosa.feature.mfcc` 関数を利用して計算し，時間平均をとった結果をCSVファイルとして保存する．"
      ],
      "metadata": {
        "id": "64wEhTscpCkT"
      }
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 318
        },
        "id": "uB8Ga2ch2BXf",
        "outputId": "1f56ffd5-576c-4fcd-8e1f-693f753622dd"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "(9997, 21)\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "       mfcc_0     mfcc_1    mfcc_2     mfcc_3    mfcc_4     mfcc_5    mfcc_6  \\\n",
              "0 -406.077911  69.616173  5.507440  14.743916  4.551410  -9.362970 -6.035157   \n",
              "1 -434.147552  77.291031  6.699678  14.172637  5.598119  -7.924884 -4.682566   \n",
              "2 -407.764404  82.102570  3.619558   8.675559  1.685305  -8.716214 -3.717378   \n",
              "3 -446.516846  59.234222  8.543715  18.219103  7.780375  -3.081481 -4.141963   \n",
              "4 -388.319855  82.455490  3.929563  12.031418  0.774454 -11.123068 -6.059217   \n",
              "\n",
              "      mfcc_7     mfcc_8    mfcc_9  ...    mfcc_11   mfcc_12   mfcc_13  \\\n",
              "0  15.677253 -16.771376 -2.346049  ... -12.573087 -3.524272 -5.719043   \n",
              "1  15.697062 -14.664041 -3.669486  ...  -9.980092 -4.588111 -3.999643   \n",
              "2  13.641570 -14.984233 -3.654552  ... -10.859071 -3.957657 -5.204740   \n",
              "3  13.044106 -10.602309 -0.732245  ...  -7.666026 -1.879559 -1.844411   \n",
              "4  15.605132 -19.640221  0.520216  ... -12.775943 -3.546141 -4.827557   \n",
              "\n",
              "     mfcc_14   mfcc_15   mfcc_16   mfcc_17   mfcc_18   mfcc_19  target  \n",
              "0 -18.324978 -2.818250 -5.371923 -3.069093 -7.313135 -2.188849       1  \n",
              "1 -16.731550 -2.347479 -4.715246  1.224066 -6.574102 -2.369225       1  \n",
              "2 -18.052351 -5.586807 -4.321259 -1.929936 -7.707550 -2.960203       1  \n",
              "3 -16.628784 -2.397813 -3.187749  0.473709 -8.178372 -2.779823       1  \n",
              "4 -18.637909 -2.985661 -7.504961  0.763457 -8.740715 -3.849721       1  \n",
              "\n",
              "[5 rows x 21 columns]"
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-69e9a09b-f84c-45a2-bfe7-8a736086508b\">\n",
              "    <div class=\"colab-df-container\">\n",
              "      <div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>mfcc_0</th>\n",
              "      <th>mfcc_1</th>\n",
              "      <th>mfcc_2</th>\n",
              "      <th>mfcc_3</th>\n",
              "      <th>mfcc_4</th>\n",
              "      <th>mfcc_5</th>\n",
              "      <th>mfcc_6</th>\n",
              "      <th>mfcc_7</th>\n",
              "      <th>mfcc_8</th>\n",
              "      <th>mfcc_9</th>\n",
              "      <th>...</th>\n",
              "      <th>mfcc_11</th>\n",
              "      <th>mfcc_12</th>\n",
              "      <th>mfcc_13</th>\n",
              "      <th>mfcc_14</th>\n",
              "      <th>mfcc_15</th>\n",
              "      <th>mfcc_16</th>\n",
              "      <th>mfcc_17</th>\n",
              "      <th>mfcc_18</th>\n",
              "      <th>mfcc_19</th>\n",
              "      <th>target</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>-406.077911</td>\n",
              "      <td>69.616173</td>\n",
              "      <td>5.507440</td>\n",
              "      <td>14.743916</td>\n",
              "      <td>4.551410</td>\n",
              "      <td>-9.362970</td>\n",
              "      <td>-6.035157</td>\n",
              "      <td>15.677253</td>\n",
              "      <td>-16.771376</td>\n",
              "      <td>-2.346049</td>\n",
              "      <td>...</td>\n",
              "      <td>-12.573087</td>\n",
              "      <td>-3.524272</td>\n",
              "      <td>-5.719043</td>\n",
              "      <td>-18.324978</td>\n",
              "      <td>-2.818250</td>\n",
              "      <td>-5.371923</td>\n",
              "      <td>-3.069093</td>\n",
              "      <td>-7.313135</td>\n",
              "      <td>-2.188849</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>-434.147552</td>\n",
              "      <td>77.291031</td>\n",
              "      <td>6.699678</td>\n",
              "      <td>14.172637</td>\n",
              "      <td>5.598119</td>\n",
              "      <td>-7.924884</td>\n",
              "      <td>-4.682566</td>\n",
              "      <td>15.697062</td>\n",
              "      <td>-14.664041</td>\n",
              "      <td>-3.669486</td>\n",
              "      <td>...</td>\n",
              "      <td>-9.980092</td>\n",
              "      <td>-4.588111</td>\n",
              "      <td>-3.999643</td>\n",
              "      <td>-16.731550</td>\n",
              "      <td>-2.347479</td>\n",
              "      <td>-4.715246</td>\n",
              "      <td>1.224066</td>\n",
              "      <td>-6.574102</td>\n",
              "      <td>-2.369225</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>-407.764404</td>\n",
              "      <td>82.102570</td>\n",
              "      <td>3.619558</td>\n",
              "      <td>8.675559</td>\n",
              "      <td>1.685305</td>\n",
              "      <td>-8.716214</td>\n",
              "      <td>-3.717378</td>\n",
              "      <td>13.641570</td>\n",
              "      <td>-14.984233</td>\n",
              "      <td>-3.654552</td>\n",
              "      <td>...</td>\n",
              "      <td>-10.859071</td>\n",
              "      <td>-3.957657</td>\n",
              "      <td>-5.204740</td>\n",
              "      <td>-18.052351</td>\n",
              "      <td>-5.586807</td>\n",
              "      <td>-4.321259</td>\n",
              "      <td>-1.929936</td>\n",
              "      <td>-7.707550</td>\n",
              "      <td>-2.960203</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>-446.516846</td>\n",
              "      <td>59.234222</td>\n",
              "      <td>8.543715</td>\n",
              "      <td>18.219103</td>\n",
              "      <td>7.780375</td>\n",
              "      <td>-3.081481</td>\n",
              "      <td>-4.141963</td>\n",
              "      <td>13.044106</td>\n",
              "      <td>-10.602309</td>\n",
              "      <td>-0.732245</td>\n",
              "      <td>...</td>\n",
              "      <td>-7.666026</td>\n",
              "      <td>-1.879559</td>\n",
              "      <td>-1.844411</td>\n",
              "      <td>-16.628784</td>\n",
              "      <td>-2.397813</td>\n",
              "      <td>-3.187749</td>\n",
              "      <td>0.473709</td>\n",
              "      <td>-8.178372</td>\n",
              "      <td>-2.779823</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>-388.319855</td>\n",
              "      <td>82.455490</td>\n",
              "      <td>3.929563</td>\n",
              "      <td>12.031418</td>\n",
              "      <td>0.774454</td>\n",
              "      <td>-11.123068</td>\n",
              "      <td>-6.059217</td>\n",
              "      <td>15.605132</td>\n",
              "      <td>-19.640221</td>\n",
              "      <td>0.520216</td>\n",
              "      <td>...</td>\n",
              "      <td>-12.775943</td>\n",
              "      <td>-3.546141</td>\n",
              "      <td>-4.827557</td>\n",
              "      <td>-18.637909</td>\n",
              "      <td>-2.985661</td>\n",
              "      <td>-7.504961</td>\n",
              "      <td>0.763457</td>\n",
              "      <td>-8.740715</td>\n",
              "      <td>-3.849721</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "<p>5 rows × 21 columns</p>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-69e9a09b-f84c-45a2-bfe7-8a736086508b')\"\n",
              "              title=\"Convert this dataframe to an interactive table.\"\n",
              "              style=\"display:none;\">\n",
              "        \n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "       width=\"24px\">\n",
              "    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
              "    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
              "  </svg>\n",
              "      </button>\n",
              "      \n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      flex-wrap:wrap;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "      <script>\n",
              "        const buttonEl =\n",
              "          document.querySelector('#df-69e9a09b-f84c-45a2-bfe7-8a736086508b button.colab-df-convert');\n",
              "        buttonEl.style.display =\n",
              "          google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "        async function convertToInteractive(key) {\n",
              "          const element = document.querySelector('#df-69e9a09b-f84c-45a2-bfe7-8a736086508b');\n",
              "          const dataTable =\n",
              "            await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                     [key], {});\n",
              "          if (!dataTable) return;\n",
              "\n",
              "          const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "            + ' to learn more about interactive tables.';\n",
              "          element.innerHTML = '';\n",
              "          dataTable['output_type'] = 'display_data';\n",
              "          await google.colab.output.renderOutput(dataTable, element);\n",
              "          const docLink = document.createElement('div');\n",
              "          docLink.innerHTML = docLinkHtml;\n",
              "          element.appendChild(docLink);\n",
              "        }\n",
              "      </script>\n",
              "    </div>\n",
              "  </div>\n",
              "  "
            ]
          },
          "metadata": {},
          "execution_count": 5
        }
      ],
      "source": [
        "import librosa\n",
        "\n",
        "X_data = []  # 特徴行列\n",
        "y_data = []  # 正解情報\n",
        "\n",
        "for speaker in range(1, 101):\n",
        "    # print(f'No.{speaker_num}') # 進捗確認\n",
        "\n",
        "    # 音声データが入っているディレクトリ名\n",
        "    dir_name = base_path + f'jvs_ver1/jvs{str(speaker).zfill(3)}/parallel100/wav24kHz16bit'\n",
        "    for file_name in os.listdir(dir_name):\n",
        "         # 音声ファイルへのパス\n",
        "        file_path = os.path.join(dir_name, file_name)\n",
        "        # 音声ファイルを読み込む\n",
        "        y, sr = librosa.load(file_path)\n",
        "\n",
        "        mfcc = librosa.feature.mfcc(y, sr) # MFCCを算出\n",
        "        mfcc = np.average(mfcc, axis=1) # 時間平均を取る\n",
        "        mfcc = mfcc.flatten()\n",
        "        mfcc = mfcc.tolist()\n",
        "        X_data.append(mfcc)\n",
        "        y_data.append(speaker)\n",
        "\n",
        "X = pd.DataFrame(X_data, columns=[f'mfcc_{n}' for n in range(0, 20)])\n",
        "y = pd.DataFrame({'target': y_data})\n",
        "\n",
        "df = pd.concat([X, y], axis=1)\n",
        "df.to_csv('mfcc.csv', index=False) # CSVで保存\n",
        "\n",
        "print(df.shape)\n",
        "df.head()"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "全データ数は $100 \\times 100 = 10^4$ だと思っていたが， $9997$ となっている．  \n",
        "元データを確認すると，いくつか音声データが欠けていることが判明した．"
      ],
      "metadata": {
        "id": "bWUMtdJ0cpdQ"
      }
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "UE9wmQC7UgAF"
      },
      "source": [
        "## 学習と評価"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "大学の講義では scikit-learn と TensorFlow を主に扱ったが，ここでは文献[2]を参考にして [PyCaret](https://pycaret.gitbook.io/docs/) というライブラリを使ってみようと思う．"
      ],
      "metadata": {
        "id": "hThDet4Hijfj"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "!pip install pycaret"
      ],
      "metadata": {
        "id": "4IkFppMf-GFd"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "from pycaret.classification import *"
      ],
      "metadata": {
        "id": "YXWvKnXsjl7L"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "保存したCSVファイルからMFCCを読み込む．"
      ],
      "metadata": {
        "id": "lwsNcFJRTAL1"
      }
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 206
        },
        "id": "L_grtxYG8dOJ",
        "outputId": "59f529dc-b511-44cc-fb90-2f7b943081d4"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "      mfcc_8    mfcc_9    mfcc_10    mfcc_11   mfcc_12   mfcc_13    mfcc_14  \\\n",
              "0 -16.771376 -2.346047 -12.703731 -12.573085 -3.524272 -5.719041 -18.324978   \n",
              "1 -14.664041 -3.669484 -11.074874  -9.980092 -4.588109 -3.999642 -16.731550   \n",
              "2 -14.984233 -3.654551 -11.442192 -10.859071 -3.957658 -5.204739 -18.052351   \n",
              "3 -10.602308 -0.732244 -11.285127  -7.666025 -1.879558 -1.844411 -16.628784   \n",
              "4 -19.640219  0.520218  -9.907987 -12.775942 -3.546140 -4.827555 -18.637911   \n",
              "\n",
              "    mfcc_15   mfcc_16   mfcc_17   mfcc_18   mfcc_19  target  \n",
              "0 -2.818249 -5.371924 -3.069092 -7.313135 -2.188848       0  \n",
              "1 -2.347478 -4.715246  1.224067 -6.574103 -2.369224       0  \n",
              "2 -5.586806 -4.321259 -1.929936 -7.707550 -2.960202       0  \n",
              "3 -2.397813 -3.187750  0.473709 -8.178372 -2.779822       0  \n",
              "4 -2.985659 -7.504962  0.763458 -8.740716 -3.849721       0  "
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-437af18a-a202-45a8-89e4-6bfa0557b604\">\n",
              "    <div class=\"colab-df-container\">\n",
              "      <div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>mfcc_8</th>\n",
              "      <th>mfcc_9</th>\n",
              "      <th>mfcc_10</th>\n",
              "      <th>mfcc_11</th>\n",
              "      <th>mfcc_12</th>\n",
              "      <th>mfcc_13</th>\n",
              "      <th>mfcc_14</th>\n",
              "      <th>mfcc_15</th>\n",
              "      <th>mfcc_16</th>\n",
              "      <th>mfcc_17</th>\n",
              "      <th>mfcc_18</th>\n",
              "      <th>mfcc_19</th>\n",
              "      <th>target</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>-16.771376</td>\n",
              "      <td>-2.346047</td>\n",
              "      <td>-12.703731</td>\n",
              "      <td>-12.573085</td>\n",
              "      <td>-3.524272</td>\n",
              "      <td>-5.719041</td>\n",
              "      <td>-18.324978</td>\n",
              "      <td>-2.818249</td>\n",
              "      <td>-5.371924</td>\n",
              "      <td>-3.069092</td>\n",
              "      <td>-7.313135</td>\n",
              "      <td>-2.188848</td>\n",
              "      <td>0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>-14.664041</td>\n",
              "      <td>-3.669484</td>\n",
              "      <td>-11.074874</td>\n",
              "      <td>-9.980092</td>\n",
              "      <td>-4.588109</td>\n",
              "      <td>-3.999642</td>\n",
              "      <td>-16.731550</td>\n",
              "      <td>-2.347478</td>\n",
              "      <td>-4.715246</td>\n",
              "      <td>1.224067</td>\n",
              "      <td>-6.574103</td>\n",
              "      <td>-2.369224</td>\n",
              "      <td>0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>-14.984233</td>\n",
              "      <td>-3.654551</td>\n",
              "      <td>-11.442192</td>\n",
              "      <td>-10.859071</td>\n",
              "      <td>-3.957658</td>\n",
              "      <td>-5.204739</td>\n",
              "      <td>-18.052351</td>\n",
              "      <td>-5.586806</td>\n",
              "      <td>-4.321259</td>\n",
              "      <td>-1.929936</td>\n",
              "      <td>-7.707550</td>\n",
              "      <td>-2.960202</td>\n",
              "      <td>0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>-10.602308</td>\n",
              "      <td>-0.732244</td>\n",
              "      <td>-11.285127</td>\n",
              "      <td>-7.666025</td>\n",
              "      <td>-1.879558</td>\n",
              "      <td>-1.844411</td>\n",
              "      <td>-16.628784</td>\n",
              "      <td>-2.397813</td>\n",
              "      <td>-3.187750</td>\n",
              "      <td>0.473709</td>\n",
              "      <td>-8.178372</td>\n",
              "      <td>-2.779822</td>\n",
              "      <td>0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>-19.640219</td>\n",
              "      <td>0.520218</td>\n",
              "      <td>-9.907987</td>\n",
              "      <td>-12.775942</td>\n",
              "      <td>-3.546140</td>\n",
              "      <td>-4.827555</td>\n",
              "      <td>-18.637911</td>\n",
              "      <td>-2.985659</td>\n",
              "      <td>-7.504962</td>\n",
              "      <td>0.763458</td>\n",
              "      <td>-8.740716</td>\n",
              "      <td>-3.849721</td>\n",
              "      <td>0</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-437af18a-a202-45a8-89e4-6bfa0557b604')\"\n",
              "              title=\"Convert this dataframe to an interactive table.\"\n",
              "              style=\"display:none;\">\n",
              "        \n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "       width=\"24px\">\n",
              "    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
              "    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
              "  </svg>\n",
              "      </button>\n",
              "      \n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      flex-wrap:wrap;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "      <script>\n",
              "        const buttonEl =\n",
              "          document.querySelector('#df-437af18a-a202-45a8-89e4-6bfa0557b604 button.colab-df-convert');\n",
              "        buttonEl.style.display =\n",
              "          google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "        async function convertToInteractive(key) {\n",
              "          const element = document.querySelector('#df-437af18a-a202-45a8-89e4-6bfa0557b604');\n",
              "          const dataTable =\n",
              "            await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                     [key], {});\n",
              "          if (!dataTable) return;\n",
              "\n",
              "          const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "            + ' to learn more about interactive tables.';\n",
              "          element.innerHTML = '';\n",
              "          dataTable['output_type'] = 'display_data';\n",
              "          await google.colab.output.renderOutput(dataTable, element);\n",
              "          const docLink = document.createElement('div');\n",
              "          docLink.innerHTML = docLinkHtml;\n",
              "          element.appendChild(docLink);\n",
              "        }\n",
              "      </script>\n",
              "    </div>\n",
              "  </div>\n",
              "  "
            ]
          },
          "metadata": {},
          "execution_count": 11
        }
      ],
      "source": [
        "data = pd.read_csv('./mfcc.csv') # 保存したCSVファイルへのパス\n",
        "\n",
        "# data = data.iloc[:, [*list(range(12)), 20]] # 低次元\n",
        "data = data.iloc[:, 8:] # 高次元\n",
        "\n",
        "data['target'] = data['target'] - 1 # クラスラベルを1ずらす\n",
        "data.head()"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "学習の設定を行う．"
      ],
      "metadata": {
        "id": "90QfN9hpTN40"
      }
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "oecW_hh97qE-"
      },
      "outputs": [],
      "source": [
        "setup(\n",
        "    data=data, target='target',\n",
        "    train_size=0.7,\n",
        "    normalize=True,\n",
        "    normalize_method='zscore', # スケーリング手法\n",
        "    data_split_shuffle=True,   # データ分割に際してデータをシャッフルする\n",
        "    fold=5,    # 交差確認法の分割数\n",
        "    n_jobs=-1, # 全プロセッサを使用して並列実行\n",
        "    use_gpu=True, silent=True)"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "利用できるモデルの一覧を表示する．"
      ],
      "metadata": {
        "id": "yyVyAKQeeE2w"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "models()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 780
        },
        "id": "xaAQF4ZSRNrC",
        "outputId": "dfc4cfad-14ba-402a-f4be-7097ceefd553"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "INFO:logs:gpu_param set to True\n",
            "WARNING:logs:Couldn't import cuml.linear_model.LogisticRegression\n",
            "WARNING:logs:Couldn't import cuml.neighbors.KNeighborsClassifier\n",
            "WARNING:logs:Couldn't import cuml.MBSGDClassifier\n",
            "WARNING:logs:Couldn't import cuml.svm.SVC\n",
            "WARNING:logs:Couldn't import cuml.linear_model\n",
            "WARNING:logs:Couldn't import cuml.ensemble\n",
            "WARNING:logs:Wrong xgboost version. Expected xgboost>=1.1.0, got xgboost==0.90\n",
            "WARNING:logs:Couldn't import catboost.CatBoostClassifier\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "                                     Name  \\\n",
              "ID                                          \n",
              "lr                    Logistic Regression   \n",
              "knn                K Neighbors Classifier   \n",
              "nb                            Naive Bayes   \n",
              "dt               Decision Tree Classifier   \n",
              "svm                   SVM - Linear Kernel   \n",
              "rbfsvm                SVM - Radial Kernel   \n",
              "gpc           Gaussian Process Classifier   \n",
              "mlp                        MLP Classifier   \n",
              "ridge                    Ridge Classifier   \n",
              "rf               Random Forest Classifier   \n",
              "qda       Quadratic Discriminant Analysis   \n",
              "ada                  Ada Boost Classifier   \n",
              "gbc          Gradient Boosting Classifier   \n",
              "lda          Linear Discriminant Analysis   \n",
              "et                 Extra Trees Classifier   \n",
              "lightgbm  Light Gradient Boosting Machine   \n",
              "dummy                    Dummy Classifier   \n",
              "\n",
              "                                                  Reference  Turbo  \n",
              "ID                                                                  \n",
              "lr        sklearn.linear_model._logistic.LogisticRegression   True  \n",
              "knn       sklearn.neighbors._classification.KNeighborsCl...   True  \n",
              "nb                           sklearn.naive_bayes.GaussianNB   True  \n",
              "dt             sklearn.tree._classes.DecisionTreeClassifier   True  \n",
              "svm       sklearn.linear_model._stochastic_gradient.SGDC...   True  \n",
              "rbfsvm                             sklearn.svm._classes.SVC  False  \n",
              "gpc       sklearn.gaussian_process._gpc.GaussianProcessC...  False  \n",
              "mlp       sklearn.neural_network._multilayer_perceptron....  False  \n",
              "ridge           sklearn.linear_model._ridge.RidgeClassifier   True  \n",
              "rf          sklearn.ensemble._forest.RandomForestClassifier   True  \n",
              "qda       sklearn.discriminant_analysis.QuadraticDiscrim...   True  \n",
              "ada       sklearn.ensemble._weight_boosting.AdaBoostClas...   True  \n",
              "gbc         sklearn.ensemble._gb.GradientBoostingClassifier   True  \n",
              "lda       sklearn.discriminant_analysis.LinearDiscrimina...   True  \n",
              "et            sklearn.ensemble._forest.ExtraTreesClassifier   True  \n",
              "lightgbm                    lightgbm.sklearn.LGBMClassifier   True  \n",
              "dummy                         sklearn.dummy.DummyClassifier   True  "
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-a5a29582-87f4-4369-9f23-ead948623e90\">\n",
              "    <div class=\"colab-df-container\">\n",
              "      <div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>Name</th>\n",
              "      <th>Reference</th>\n",
              "      <th>Turbo</th>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>ID</th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>lr</th>\n",
              "      <td>Logistic Regression</td>\n",
              "      <td>sklearn.linear_model._logistic.LogisticRegression</td>\n",
              "      <td>True</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>knn</th>\n",
              "      <td>K Neighbors Classifier</td>\n",
              "      <td>sklearn.neighbors._classification.KNeighborsCl...</td>\n",
              "      <td>True</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>nb</th>\n",
              "      <td>Naive Bayes</td>\n",
              "      <td>sklearn.naive_bayes.GaussianNB</td>\n",
              "      <td>True</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>dt</th>\n",
              "      <td>Decision Tree Classifier</td>\n",
              "      <td>sklearn.tree._classes.DecisionTreeClassifier</td>\n",
              "      <td>True</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>svm</th>\n",
              "      <td>SVM - Linear Kernel</td>\n",
              "      <td>sklearn.linear_model._stochastic_gradient.SGDC...</td>\n",
              "      <td>True</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>rbfsvm</th>\n",
              "      <td>SVM - Radial Kernel</td>\n",
              "      <td>sklearn.svm._classes.SVC</td>\n",
              "      <td>False</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>gpc</th>\n",
              "      <td>Gaussian Process Classifier</td>\n",
              "      <td>sklearn.gaussian_process._gpc.GaussianProcessC...</td>\n",
              "      <td>False</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>mlp</th>\n",
              "      <td>MLP Classifier</td>\n",
              "      <td>sklearn.neural_network._multilayer_perceptron....</td>\n",
              "      <td>False</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>ridge</th>\n",
              "      <td>Ridge Classifier</td>\n",
              "      <td>sklearn.linear_model._ridge.RidgeClassifier</td>\n",
              "      <td>True</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>rf</th>\n",
              "      <td>Random Forest Classifier</td>\n",
              "      <td>sklearn.ensemble._forest.RandomForestClassifier</td>\n",
              "      <td>True</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>qda</th>\n",
              "      <td>Quadratic Discriminant Analysis</td>\n",
              "      <td>sklearn.discriminant_analysis.QuadraticDiscrim...</td>\n",
              "      <td>True</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>ada</th>\n",
              "      <td>Ada Boost Classifier</td>\n",
              "      <td>sklearn.ensemble._weight_boosting.AdaBoostClas...</td>\n",
              "      <td>True</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>gbc</th>\n",
              "      <td>Gradient Boosting Classifier</td>\n",
              "      <td>sklearn.ensemble._gb.GradientBoostingClassifier</td>\n",
              "      <td>True</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>lda</th>\n",
              "      <td>Linear Discriminant Analysis</td>\n",
              "      <td>sklearn.discriminant_analysis.LinearDiscrimina...</td>\n",
              "      <td>True</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>et</th>\n",
              "      <td>Extra Trees Classifier</td>\n",
              "      <td>sklearn.ensemble._forest.ExtraTreesClassifier</td>\n",
              "      <td>True</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>lightgbm</th>\n",
              "      <td>Light Gradient Boosting Machine</td>\n",
              "      <td>lightgbm.sklearn.LGBMClassifier</td>\n",
              "      <td>True</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>dummy</th>\n",
              "      <td>Dummy Classifier</td>\n",
              "      <td>sklearn.dummy.DummyClassifier</td>\n",
              "      <td>True</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-a5a29582-87f4-4369-9f23-ead948623e90')\"\n",
              "              title=\"Convert this dataframe to an interactive table.\"\n",
              "              style=\"display:none;\">\n",
              "        \n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "       width=\"24px\">\n",
              "    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
              "    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
              "  </svg>\n",
              "      </button>\n",
              "      \n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      flex-wrap:wrap;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "      <script>\n",
              "        const buttonEl =\n",
              "          document.querySelector('#df-a5a29582-87f4-4369-9f23-ead948623e90 button.colab-df-convert');\n",
              "        buttonEl.style.display =\n",
              "          google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "        async function convertToInteractive(key) {\n",
              "          const element = document.querySelector('#df-a5a29582-87f4-4369-9f23-ead948623e90');\n",
              "          const dataTable =\n",
              "            await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                     [key], {});\n",
              "          if (!dataTable) return;\n",
              "\n",
              "          const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "            + ' to learn more about interactive tables.';\n",
              "          element.innerHTML = '';\n",
              "          dataTable['output_type'] = 'display_data';\n",
              "          await google.colab.output.renderOutput(dataTable, element);\n",
              "          const docLink = document.createElement('div');\n",
              "          docLink.innerHTML = docLinkHtml;\n",
              "          element.appendChild(docLink);\n",
              "        }\n",
              "      </script>\n",
              "    </div>\n",
              "  </div>\n",
              "  "
            ]
          },
          "metadata": {},
          "execution_count": 9
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "極端に学習が遅いモデルは除外して性能を比較する．"
      ],
      "metadata": {
        "id": "Jx5xEaypTSFc"
      }
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 567,
          "referenced_widgets": [
            "0fbcc8c8323344f68c5514621b16a3f9",
            "8969511da9ad415599e0c7902a0cda2a",
            "cf5f6a1e672f4814a2bd2d8aad9e5348"
          ]
        },
        "id": "N0xipCgA8aBp",
        "outputId": "9e0c1cad-915e-4b7d-e3d3-d603b4827508"
      },
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "                                    Model  Accuracy     AUC  Recall   Prec.  \\\n",
              "qda       Quadratic Discriminant Analysis    0.9767  0.9999  0.9767  0.9789   \n",
              "knn                K Neighbors Classifier    0.9608  0.9960  0.9601  0.9634   \n",
              "et                 Extra Trees Classifier    0.9568  0.9995  0.9561  0.9594   \n",
              "lr                    Logistic Regression    0.9457  0.9996  0.9453  0.9484   \n",
              "rf               Random Forest Classifier    0.9413  0.9993  0.9410  0.9440   \n",
              "lda          Linear Discriminant Analysis    0.9345  0.9995  0.9344  0.9388   \n",
              "nb                            Naive Bayes    0.9138  0.9990  0.9133  0.9199   \n",
              "lightgbm  Light Gradient Boosting Machine    0.8872  0.9967  0.8861  0.8932   \n",
              "svm                   SVM - Linear Kernel    0.7485  0.0000  0.7473  0.9058   \n",
              "dt               Decision Tree Classifier    0.6872  0.8420  0.6862  0.7032   \n",
              "ridge                    Ridge Classifier    0.6576  0.0000  0.6543  0.7166   \n",
              "ada                  Ada Boost Classifier    0.0616  0.7010  0.0610  0.0415   \n",
              "dummy                    Dummy Classifier    0.0114  0.5000  0.0100  0.0001   \n",
              "\n",
              "              F1   Kappa     MCC  TT (Sec)  \n",
              "qda       0.9767  0.9765  0.9765     0.250  \n",
              "knn       0.9603  0.9604  0.9605     1.172  \n",
              "et        0.9564  0.9564  0.9564     6.866  \n",
              "lr        0.9453  0.9451  0.9452     2.366  \n",
              "rf        0.9407  0.9407  0.9407     9.424  \n",
              "lda       0.9342  0.9339  0.9339     0.564  \n",
              "nb        0.9142  0.9129  0.9130     0.236  \n",
              "lightgbm  0.8870  0.8861  0.8862    29.918  \n",
              "svm       0.7956  0.7459  0.7506     0.764  \n",
              "dt        0.6876  0.6840  0.6841     0.580  \n",
              "ridge     0.6174  0.6541  0.6565     0.066  \n",
              "ada       0.0342  0.0515  0.0626     2.472  \n",
              "dummy     0.0003  0.0000  0.0000     0.180  "
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-a3de4441-690d-42e7-8fd0-12f158673575\">\n",
              "    <div class=\"colab-df-container\">\n",
              "      <div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>Model</th>\n",
              "      <th>Accuracy</th>\n",
              "      <th>AUC</th>\n",
              "      <th>Recall</th>\n",
              "      <th>Prec.</th>\n",
              "      <th>F1</th>\n",
              "      <th>Kappa</th>\n",
              "      <th>MCC</th>\n",
              "      <th>TT (Sec)</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>qda</th>\n",
              "      <td>Quadratic Discriminant Analysis</td>\n",
              "      <td>0.9767</td>\n",
              "      <td>0.9999</td>\n",
              "      <td>0.9767</td>\n",
              "      <td>0.9789</td>\n",
              "      <td>0.9767</td>\n",
              "      <td>0.9765</td>\n",
              "      <td>0.9765</td>\n",
              "      <td>0.250</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>knn</th>\n",
              "      <td>K Neighbors Classifier</td>\n",
              "      <td>0.9608</td>\n",
              "      <td>0.9960</td>\n",
              "      <td>0.9601</td>\n",
              "      <td>0.9634</td>\n",
              "      <td>0.9603</td>\n",
              "      <td>0.9604</td>\n",
              "      <td>0.9605</td>\n",
              "      <td>1.172</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>et</th>\n",
              "      <td>Extra Trees Classifier</td>\n",
              "      <td>0.9568</td>\n",
              "      <td>0.9995</td>\n",
              "      <td>0.9561</td>\n",
              "      <td>0.9594</td>\n",
              "      <td>0.9564</td>\n",
              "      <td>0.9564</td>\n",
              "      <td>0.9564</td>\n",
              "      <td>6.866</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>lr</th>\n",
              "      <td>Logistic Regression</td>\n",
              "      <td>0.9457</td>\n",
              "      <td>0.9996</td>\n",
              "      <td>0.9453</td>\n",
              "      <td>0.9484</td>\n",
              "      <td>0.9453</td>\n",
              "      <td>0.9451</td>\n",
              "      <td>0.9452</td>\n",
              "      <td>2.366</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>rf</th>\n",
              "      <td>Random Forest Classifier</td>\n",
              "      <td>0.9413</td>\n",
              "      <td>0.9993</td>\n",
              "      <td>0.9410</td>\n",
              "      <td>0.9440</td>\n",
              "      <td>0.9407</td>\n",
              "      <td>0.9407</td>\n",
              "      <td>0.9407</td>\n",
              "      <td>9.424</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>lda</th>\n",
              "      <td>Linear Discriminant Analysis</td>\n",
              "      <td>0.9345</td>\n",
              "      <td>0.9995</td>\n",
              "      <td>0.9344</td>\n",
              "      <td>0.9388</td>\n",
              "      <td>0.9342</td>\n",
              "      <td>0.9339</td>\n",
              "      <td>0.9339</td>\n",
              "      <td>0.564</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>nb</th>\n",
              "      <td>Naive Bayes</td>\n",
              "      <td>0.9138</td>\n",
              "      <td>0.9990</td>\n",
              "      <td>0.9133</td>\n",
              "      <td>0.9199</td>\n",
              "      <td>0.9142</td>\n",
              "      <td>0.9129</td>\n",
              "      <td>0.9130</td>\n",
              "      <td>0.236</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>lightgbm</th>\n",
              "      <td>Light Gradient Boosting Machine</td>\n",
              "      <td>0.8872</td>\n",
              "      <td>0.9967</td>\n",
              "      <td>0.8861</td>\n",
              "      <td>0.8932</td>\n",
              "      <td>0.8870</td>\n",
              "      <td>0.8861</td>\n",
              "      <td>0.8862</td>\n",
              "      <td>29.918</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>svm</th>\n",
              "      <td>SVM - Linear Kernel</td>\n",
              "      <td>0.7485</td>\n",
              "      <td>0.0000</td>\n",
              "      <td>0.7473</td>\n",
              "      <td>0.9058</td>\n",
              "      <td>0.7956</td>\n",
              "      <td>0.7459</td>\n",
              "      <td>0.7506</td>\n",
              "      <td>0.764</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>dt</th>\n",
              "      <td>Decision Tree Classifier</td>\n",
              "      <td>0.6872</td>\n",
              "      <td>0.8420</td>\n",
              "      <td>0.6862</td>\n",
              "      <td>0.7032</td>\n",
              "      <td>0.6876</td>\n",
              "      <td>0.6840</td>\n",
              "      <td>0.6841</td>\n",
              "      <td>0.580</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>ridge</th>\n",
              "      <td>Ridge Classifier</td>\n",
              "      <td>0.6576</td>\n",
              "      <td>0.0000</td>\n",
              "      <td>0.6543</td>\n",
              "      <td>0.7166</td>\n",
              "      <td>0.6174</td>\n",
              "      <td>0.6541</td>\n",
              "      <td>0.6565</td>\n",
              "      <td>0.066</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>ada</th>\n",
              "      <td>Ada Boost Classifier</td>\n",
              "      <td>0.0616</td>\n",
              "      <td>0.7010</td>\n",
              "      <td>0.0610</td>\n",
              "      <td>0.0415</td>\n",
              "      <td>0.0342</td>\n",
              "      <td>0.0515</td>\n",
              "      <td>0.0626</td>\n",
              "      <td>2.472</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>dummy</th>\n",
              "      <td>Dummy Classifier</td>\n",
              "      <td>0.0114</td>\n",
              "      <td>0.5000</td>\n",
              "      <td>0.0100</td>\n",
              "      <td>0.0001</td>\n",
              "      <td>0.0003</td>\n",
              "      <td>0.0000</td>\n",
              "      <td>0.0000</td>\n",
              "      <td>0.180</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-a3de4441-690d-42e7-8fd0-12f158673575')\"\n",
              "              title=\"Convert this dataframe to an interactive table.\"\n",
              "              style=\"display:none;\">\n",
              "        \n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "       width=\"24px\">\n",
              "    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
              "    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
              "  </svg>\n",
              "      </button>\n",
              "      \n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      flex-wrap:wrap;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "      <script>\n",
              "        const buttonEl =\n",
              "          document.querySelector('#df-a3de4441-690d-42e7-8fd0-12f158673575 button.colab-df-convert');\n",
              "        buttonEl.style.display =\n",
              "          google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "        async function convertToInteractive(key) {\n",
              "          const element = document.querySelector('#df-a3de4441-690d-42e7-8fd0-12f158673575');\n",
              "          const dataTable =\n",
              "            await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                     [key], {});\n",
              "          if (!dataTable) return;\n",
              "\n",
              "          const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "            + ' to learn more about interactive tables.';\n",
              "          element.innerHTML = '';\n",
              "          dataTable['output_type'] = 'display_data';\n",
              "          await google.colab.output.renderOutput(dataTable, element);\n",
              "          const docLink = document.createElement('div');\n",
              "          docLink.innerHTML = docLinkHtml;\n",
              "          element.appendChild(docLink);\n",
              "        }\n",
              "      </script>\n",
              "    </div>\n",
              "  </div>\n",
              "  "
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "INFO:logs:create_model_container: 13\n",
            "INFO:logs:master_model_container: 13\n",
            "INFO:logs:display_container: 2\n",
            "INFO:logs:QuadraticDiscriminantAnalysis(priors=None, reg_param=0.0,\n",
            "                              store_covariance=False, tol=0.0001)\n",
            "INFO:logs:compare_models() succesfully completed......................................\n"
          ]
        }
      ],
      "source": [
        "best = compare_models(exclude=['gbc'])"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "`gbc` 以外の全てのモデルが評価対象に含まれてないのは何故だろう．"
      ],
      "metadata": {
        "id": "DaD2E0P5gmZv"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "直前のコードセルの出力は，各モデルのデフォルトのパラメータを用いて学習させた結果である．PyCaret ではここからパラメータの調整も関数1つでできるようになっている．恐るべし．"
      ],
      "metadata": {
        "id": "at4fIrxUfz3l"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "tuned = tune_model(best)\n",
        "predict_model(tuned)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000,
          "referenced_widgets": [
            "610467b19ad846f88f4b710f210766a4",
            "c8f9271961c14406a4555f393efbe555",
            "553005b76fff4f8b9eec8b2dd338a6a0"
          ]
        },
        "id": "_tFQW7sEfcKh",
        "outputId": "1fdadaf2-43f5-4425-c7e8-8ca5fac81705"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "      Accuracy     AUC  Recall   Prec.      F1   Kappa     MCC\n",
              "Fold                                                          \n",
              "0       0.9679  0.9999  0.9670  0.9704  0.9677  0.9675  0.9676\n",
              "1       0.9643  0.9999  0.9636  0.9668  0.9641  0.9639  0.9640\n",
              "2       0.9678  0.9999  0.9675  0.9699  0.9675  0.9675  0.9675\n",
              "3       0.9600  0.9998  0.9595  0.9624  0.9593  0.9596  0.9596\n",
              "4       0.9650  0.9999  0.9648  0.9672  0.9647  0.9646  0.9646\n",
              "Mean    0.9650  0.9999  0.9645  0.9673  0.9647  0.9646  0.9647\n",
              "Std     0.0029  0.0000  0.0029  0.0029  0.0031  0.0029  0.0029"
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-ed96312f-4122-46c7-a31f-8fcb5b51d4df\">\n",
              "    <div class=\"colab-df-container\">\n",
              "      <div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>Accuracy</th>\n",
              "      <th>AUC</th>\n",
              "      <th>Recall</th>\n",
              "      <th>Prec.</th>\n",
              "      <th>F1</th>\n",
              "      <th>Kappa</th>\n",
              "      <th>MCC</th>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>Fold</th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>0.9679</td>\n",
              "      <td>0.9999</td>\n",
              "      <td>0.9670</td>\n",
              "      <td>0.9704</td>\n",
              "      <td>0.9677</td>\n",
              "      <td>0.9675</td>\n",
              "      <td>0.9676</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>0.9643</td>\n",
              "      <td>0.9999</td>\n",
              "      <td>0.9636</td>\n",
              "      <td>0.9668</td>\n",
              "      <td>0.9641</td>\n",
              "      <td>0.9639</td>\n",
              "      <td>0.9640</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>0.9678</td>\n",
              "      <td>0.9999</td>\n",
              "      <td>0.9675</td>\n",
              "      <td>0.9699</td>\n",
              "      <td>0.9675</td>\n",
              "      <td>0.9675</td>\n",
              "      <td>0.9675</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>0.9600</td>\n",
              "      <td>0.9998</td>\n",
              "      <td>0.9595</td>\n",
              "      <td>0.9624</td>\n",
              "      <td>0.9593</td>\n",
              "      <td>0.9596</td>\n",
              "      <td>0.9596</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>0.9650</td>\n",
              "      <td>0.9999</td>\n",
              "      <td>0.9648</td>\n",
              "      <td>0.9672</td>\n",
              "      <td>0.9647</td>\n",
              "      <td>0.9646</td>\n",
              "      <td>0.9646</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>Mean</th>\n",
              "      <td>0.9650</td>\n",
              "      <td>0.9999</td>\n",
              "      <td>0.9645</td>\n",
              "      <td>0.9673</td>\n",
              "      <td>0.9647</td>\n",
              "      <td>0.9646</td>\n",
              "      <td>0.9647</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>Std</th>\n",
              "      <td>0.0029</td>\n",
              "      <td>0.0000</td>\n",
              "      <td>0.0029</td>\n",
              "      <td>0.0029</td>\n",
              "      <td>0.0031</td>\n",
              "      <td>0.0029</td>\n",
              "      <td>0.0029</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-ed96312f-4122-46c7-a31f-8fcb5b51d4df')\"\n",
              "              title=\"Convert this dataframe to an interactive table.\"\n",
              "              style=\"display:none;\">\n",
              "        \n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "       width=\"24px\">\n",
              "    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
              "    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
              "  </svg>\n",
              "      </button>\n",
              "      \n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      flex-wrap:wrap;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "      <script>\n",
              "        const buttonEl =\n",
              "          document.querySelector('#df-ed96312f-4122-46c7-a31f-8fcb5b51d4df button.colab-df-convert');\n",
              "        buttonEl.style.display =\n",
              "          google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "        async function convertToInteractive(key) {\n",
              "          const element = document.querySelector('#df-ed96312f-4122-46c7-a31f-8fcb5b51d4df');\n",
              "          const dataTable =\n",
              "            await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                     [key], {});\n",
              "          if (!dataTable) return;\n",
              "\n",
              "          const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "            + ' to learn more about interactive tables.';\n",
              "          element.innerHTML = '';\n",
              "          dataTable['output_type'] = 'display_data';\n",
              "          await google.colab.output.renderOutput(dataTable, element);\n",
              "          const docLink = document.createElement('div');\n",
              "          docLink.innerHTML = docLinkHtml;\n",
              "          element.appendChild(docLink);\n",
              "        }\n",
              "      </script>\n",
              "    </div>\n",
              "  </div>\n",
              "  "
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "INFO:logs:create_model_container: 14\n",
            "INFO:logs:master_model_container: 14\n",
            "INFO:logs:display_container: 3\n",
            "INFO:logs:QuadraticDiscriminantAnalysis(priors=None, reg_param=0.26,\n",
            "                              store_covariance=False, tol=0.0001)\n",
            "INFO:logs:tune_model() succesfully completed......................................\n",
            "INFO:logs:Initializing predict_model()\n",
            "INFO:logs:predict_model(estimator=QuadraticDiscriminantAnalysis(priors=None, reg_param=0.26,\n",
            "                              store_covariance=False, tol=0.0001), probability_threshold=None, encoded_labels=False, drift_report=False, raw_score=False, round=4, verbose=True, ml_usecase=MLUsecase.CLASSIFICATION, display=None, drift_kwargs=None)\n",
            "INFO:logs:Checking exceptions\n",
            "INFO:logs:Preloading libraries\n",
            "INFO:logs:Preparing display monitor\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "                             Model  Accuracy     AUC  Recall   Prec.     F1  \\\n",
              "0  Quadratic Discriminant Analysis    0.9633  0.9995   0.966  0.9663  0.963   \n",
              "\n",
              "   Kappa    MCC  \n",
              "0  0.963  0.963  "
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-5e933ab0-968c-465b-bdfe-2dc0f3114ce3\">\n",
              "    <div class=\"colab-df-container\">\n",
              "      <div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>Model</th>\n",
              "      <th>Accuracy</th>\n",
              "      <th>AUC</th>\n",
              "      <th>Recall</th>\n",
              "      <th>Prec.</th>\n",
              "      <th>F1</th>\n",
              "      <th>Kappa</th>\n",
              "      <th>MCC</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>Quadratic Discriminant Analysis</td>\n",
              "      <td>0.9633</td>\n",
              "      <td>0.9995</td>\n",
              "      <td>0.966</td>\n",
              "      <td>0.9663</td>\n",
              "      <td>0.963</td>\n",
              "      <td>0.963</td>\n",
              "      <td>0.963</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-5e933ab0-968c-465b-bdfe-2dc0f3114ce3')\"\n",
              "              title=\"Convert this dataframe to an interactive table.\"\n",
              "              style=\"display:none;\">\n",
              "        \n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "       width=\"24px\">\n",
              "    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
              "    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
              "  </svg>\n",
              "      </button>\n",
              "      \n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      flex-wrap:wrap;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "      <script>\n",
              "        const buttonEl =\n",
              "          document.querySelector('#df-5e933ab0-968c-465b-bdfe-2dc0f3114ce3 button.colab-df-convert');\n",
              "        buttonEl.style.display =\n",
              "          google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "        async function convertToInteractive(key) {\n",
              "          const element = document.querySelector('#df-5e933ab0-968c-465b-bdfe-2dc0f3114ce3');\n",
              "          const dataTable =\n",
              "            await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                     [key], {});\n",
              "          if (!dataTable) return;\n",
              "\n",
              "          const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "            + ' to learn more about interactive tables.';\n",
              "          element.innerHTML = '';\n",
              "          dataTable['output_type'] = 'display_data';\n",
              "          await google.colab.output.renderOutput(dataTable, element);\n",
              "          const docLink = document.createElement('div');\n",
              "          docLink.innerHTML = docLinkHtml;\n",
              "          element.appendChild(docLink);\n",
              "        }\n",
              "      </script>\n",
              "    </div>\n",
              "  </div>\n",
              "  "
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "        mfcc_8    mfcc_9   mfcc_10   mfcc_11   mfcc_12   mfcc_13   mfcc_14  \\\n",
              "0    -1.791273 -0.813080  0.540968 -0.172749 -0.941744  1.706620 -1.363003   \n",
              "1    -1.424081 -0.161605 -1.502280  1.246076 -0.946695 -1.071474  0.907517   \n",
              "2    -0.603853  0.409088 -1.275670  1.415898 -0.322849  1.390981 -0.669469   \n",
              "3     1.125904  0.290464 -0.961848  1.078983 -1.295378 -0.017969 -0.462431   \n",
              "4     0.844751  1.179189  0.367359  0.954606 -0.345100 -1.323318  1.835905   \n",
              "...        ...       ...       ...       ...       ...       ...       ...   \n",
              "2995  0.366480  1.215562 -1.550293 -0.433060 -1.497672  2.020582 -0.206214   \n",
              "2996  0.186567 -0.324863  1.622369 -1.187052 -0.289645 -1.102727  1.032621   \n",
              "2997  1.476891 -0.932415  1.165538 -0.557417  0.929140 -0.348897  0.053719   \n",
              "2998 -1.182471 -0.355979 -0.295684 -1.605903  0.384712 -1.051158 -3.144795   \n",
              "2999  1.314170  0.532667  1.799296  1.225391  1.439906  1.146070  0.046036   \n",
              "\n",
              "       mfcc_15   mfcc_16   mfcc_17   mfcc_18   mfcc_19  target  Label   Score  \n",
              "0     0.390721 -0.146744  0.698333  2.506359  2.233332      13     13  1.0000  \n",
              "1    -1.852941 -0.422839  0.937872 -3.528243  1.062655      76     76  1.0000  \n",
              "2     0.030509  0.064247 -0.928644  0.408001 -1.100312      64     64  0.9803  \n",
              "3    -0.225192  0.644663 -0.137468 -0.756707  1.048512      15     15  0.7758  \n",
              "4     0.160658 -0.639797  0.666306  0.247593 -0.393702      99     99  0.9525  \n",
              "...        ...       ...       ...       ...       ...     ...    ...     ...  \n",
              "2995  0.518060 -0.120394 -0.238255 -1.335594 -0.191704      62     62  0.9979  \n",
              "2996 -0.482762 -1.404019  1.637034 -0.103544 -0.631369      74     74  0.9991  \n",
              "2997  0.860161 -0.539392 -0.667244  0.941765  0.069399      89     89  0.9343  \n",
              "2998 -0.082152 -0.338130 -0.014212 -0.228734 -0.423104       0      0  1.0000  \n",
              "2999  0.814284  1.397306 -0.615039  1.075020 -0.542079      83     83  0.9986  \n",
              "\n",
              "[3000 rows x 15 columns]"
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-228128ee-6d4f-4a47-9c9b-defe50673165\">\n",
              "    <div class=\"colab-df-container\">\n",
              "      <div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>mfcc_8</th>\n",
              "      <th>mfcc_9</th>\n",
              "      <th>mfcc_10</th>\n",
              "      <th>mfcc_11</th>\n",
              "      <th>mfcc_12</th>\n",
              "      <th>mfcc_13</th>\n",
              "      <th>mfcc_14</th>\n",
              "      <th>mfcc_15</th>\n",
              "      <th>mfcc_16</th>\n",
              "      <th>mfcc_17</th>\n",
              "      <th>mfcc_18</th>\n",
              "      <th>mfcc_19</th>\n",
              "      <th>target</th>\n",
              "      <th>Label</th>\n",
              "      <th>Score</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>-1.791273</td>\n",
              "      <td>-0.813080</td>\n",
              "      <td>0.540968</td>\n",
              "      <td>-0.172749</td>\n",
              "      <td>-0.941744</td>\n",
              "      <td>1.706620</td>\n",
              "      <td>-1.363003</td>\n",
              "      <td>0.390721</td>\n",
              "      <td>-0.146744</td>\n",
              "      <td>0.698333</td>\n",
              "      <td>2.506359</td>\n",
              "      <td>2.233332</td>\n",
              "      <td>13</td>\n",
              "      <td>13</td>\n",
              "      <td>1.0000</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>-1.424081</td>\n",
              "      <td>-0.161605</td>\n",
              "      <td>-1.502280</td>\n",
              "      <td>1.246076</td>\n",
              "      <td>-0.946695</td>\n",
              "      <td>-1.071474</td>\n",
              "      <td>0.907517</td>\n",
              "      <td>-1.852941</td>\n",
              "      <td>-0.422839</td>\n",
              "      <td>0.937872</td>\n",
              "      <td>-3.528243</td>\n",
              "      <td>1.062655</td>\n",
              "      <td>76</td>\n",
              "      <td>76</td>\n",
              "      <td>1.0000</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>-0.603853</td>\n",
              "      <td>0.409088</td>\n",
              "      <td>-1.275670</td>\n",
              "      <td>1.415898</td>\n",
              "      <td>-0.322849</td>\n",
              "      <td>1.390981</td>\n",
              "      <td>-0.669469</td>\n",
              "      <td>0.030509</td>\n",
              "      <td>0.064247</td>\n",
              "      <td>-0.928644</td>\n",
              "      <td>0.408001</td>\n",
              "      <td>-1.100312</td>\n",
              "      <td>64</td>\n",
              "      <td>64</td>\n",
              "      <td>0.9803</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>1.125904</td>\n",
              "      <td>0.290464</td>\n",
              "      <td>-0.961848</td>\n",
              "      <td>1.078983</td>\n",
              "      <td>-1.295378</td>\n",
              "      <td>-0.017969</td>\n",
              "      <td>-0.462431</td>\n",
              "      <td>-0.225192</td>\n",
              "      <td>0.644663</td>\n",
              "      <td>-0.137468</td>\n",
              "      <td>-0.756707</td>\n",
              "      <td>1.048512</td>\n",
              "      <td>15</td>\n",
              "      <td>15</td>\n",
              "      <td>0.7758</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>0.844751</td>\n",
              "      <td>1.179189</td>\n",
              "      <td>0.367359</td>\n",
              "      <td>0.954606</td>\n",
              "      <td>-0.345100</td>\n",
              "      <td>-1.323318</td>\n",
              "      <td>1.835905</td>\n",
              "      <td>0.160658</td>\n",
              "      <td>-0.639797</td>\n",
              "      <td>0.666306</td>\n",
              "      <td>0.247593</td>\n",
              "      <td>-0.393702</td>\n",
              "      <td>99</td>\n",
              "      <td>99</td>\n",
              "      <td>0.9525</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>...</th>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2995</th>\n",
              "      <td>0.366480</td>\n",
              "      <td>1.215562</td>\n",
              "      <td>-1.550293</td>\n",
              "      <td>-0.433060</td>\n",
              "      <td>-1.497672</td>\n",
              "      <td>2.020582</td>\n",
              "      <td>-0.206214</td>\n",
              "      <td>0.518060</td>\n",
              "      <td>-0.120394</td>\n",
              "      <td>-0.238255</td>\n",
              "      <td>-1.335594</td>\n",
              "      <td>-0.191704</td>\n",
              "      <td>62</td>\n",
              "      <td>62</td>\n",
              "      <td>0.9979</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2996</th>\n",
              "      <td>0.186567</td>\n",
              "      <td>-0.324863</td>\n",
              "      <td>1.622369</td>\n",
              "      <td>-1.187052</td>\n",
              "      <td>-0.289645</td>\n",
              "      <td>-1.102727</td>\n",
              "      <td>1.032621</td>\n",
              "      <td>-0.482762</td>\n",
              "      <td>-1.404019</td>\n",
              "      <td>1.637034</td>\n",
              "      <td>-0.103544</td>\n",
              "      <td>-0.631369</td>\n",
              "      <td>74</td>\n",
              "      <td>74</td>\n",
              "      <td>0.9991</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2997</th>\n",
              "      <td>1.476891</td>\n",
              "      <td>-0.932415</td>\n",
              "      <td>1.165538</td>\n",
              "      <td>-0.557417</td>\n",
              "      <td>0.929140</td>\n",
              "      <td>-0.348897</td>\n",
              "      <td>0.053719</td>\n",
              "      <td>0.860161</td>\n",
              "      <td>-0.539392</td>\n",
              "      <td>-0.667244</td>\n",
              "      <td>0.941765</td>\n",
              "      <td>0.069399</td>\n",
              "      <td>89</td>\n",
              "      <td>89</td>\n",
              "      <td>0.9343</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2998</th>\n",
              "      <td>-1.182471</td>\n",
              "      <td>-0.355979</td>\n",
              "      <td>-0.295684</td>\n",
              "      <td>-1.605903</td>\n",
              "      <td>0.384712</td>\n",
              "      <td>-1.051158</td>\n",
              "      <td>-3.144795</td>\n",
              "      <td>-0.082152</td>\n",
              "      <td>-0.338130</td>\n",
              "      <td>-0.014212</td>\n",
              "      <td>-0.228734</td>\n",
              "      <td>-0.423104</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1.0000</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2999</th>\n",
              "      <td>1.314170</td>\n",
              "      <td>0.532667</td>\n",
              "      <td>1.799296</td>\n",
              "      <td>1.225391</td>\n",
              "      <td>1.439906</td>\n",
              "      <td>1.146070</td>\n",
              "      <td>0.046036</td>\n",
              "      <td>0.814284</td>\n",
              "      <td>1.397306</td>\n",
              "      <td>-0.615039</td>\n",
              "      <td>1.075020</td>\n",
              "      <td>-0.542079</td>\n",
              "      <td>83</td>\n",
              "      <td>83</td>\n",
              "      <td>0.9986</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "<p>3000 rows × 15 columns</p>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-228128ee-6d4f-4a47-9c9b-defe50673165')\"\n",
              "              title=\"Convert this dataframe to an interactive table.\"\n",
              "              style=\"display:none;\">\n",
              "        \n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "       width=\"24px\">\n",
              "    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
              "    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
              "  </svg>\n",
              "      </button>\n",
              "      \n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      flex-wrap:wrap;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "      <script>\n",
              "        const buttonEl =\n",
              "          document.querySelector('#df-228128ee-6d4f-4a47-9c9b-defe50673165 button.colab-df-convert');\n",
              "        buttonEl.style.display =\n",
              "          google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "        async function convertToInteractive(key) {\n",
              "          const element = document.querySelector('#df-228128ee-6d4f-4a47-9c9b-defe50673165');\n",
              "          const dataTable =\n",
              "            await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                     [key], {});\n",
              "          if (!dataTable) return;\n",
              "\n",
              "          const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "            + ' to learn more about interactive tables.';\n",
              "          element.innerHTML = '';\n",
              "          dataTable['output_type'] = 'display_data';\n",
              "          await google.colab.output.renderOutput(dataTable, element);\n",
              "          const docLink = document.createElement('div');\n",
              "          docLink.innerHTML = docLinkHtml;\n",
              "          element.appendChild(docLink);\n",
              "        }\n",
              "      </script>\n",
              "    </div>\n",
              "  </div>\n",
              "  "
            ]
          },
          "metadata": {},
          "execution_count": 14
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "## 結果と考察"
      ],
      "metadata": {
        "id": "B7DrwgvqjB4S"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "MFCCの低次から12次元取り出した特徴量と，高次から12次元取り出した特徴量を用いた場合の正解率を以下の表にまとめる．"
      ],
      "metadata": {
        "id": "P29fMJw-7fPK"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "| モデル   | 低次元 | 高次元 |\n",
        "| -------: | :----: | :----: |\n",
        "| qda      | 0.9831 | 0.9767 |\n",
        "| lda      | 0.9754 | 0.9345 |\n",
        "| lr       | 0.9723 | 0.9457 |\n",
        "| et       | 0.9653 | 0.9508 |\n",
        "| knn      | 0.9644 | 0.9608 |\n",
        "| rf       | 0.9500 | 0.9413 |\n",
        "| nb       | 0.9487 | 0.9138 |\n",
        "| lightgbm | 0.9050 | 0.8872 |\n",
        "| svm      | 0.8139 | 0.7485 |\n",
        "| dt       | 0.7290 | 0.6872 |\n",
        "| ridge    | 0.6867 | 0.6576 |\n",
        "| ada      | 0.0762 | 0.0616 |\n",
        "| dummy    | 0.0110 | 0.0114 |\n"
      ],
      "metadata": {
        "id": "4CdN7P189Jq9"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "予想に反して，高次元よりも低次元の情報を使った方が全体的に正解率は高かった．これは，MFCCの低次元の情報にも話者認識に有効な特徴が含まれていることを意味する．\n",
        "\n",
        "音声は，肺から押し出された空気が声帯の制御によってパルス波になり，それが口と鼻から成る声道で共振して生み出される．音素の違いは，口の開き具合や舌の位置といった声道の変形によって生じるため，音声認識にはMFCCの低次元の情報が有効であることは先に述べた通りである．低次元の情報を用いても正確に話者を識別できたことから，声道の形状には個人差があり，それが共振周波数の違いとなって現れた可能性が考えられる．"
      ],
      "metadata": {
        "id": "xDKlfjxjjOcW"
      }
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "RLCkp5RU0wkV"
      },
      "source": [
        "## 参考文献"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Wkr9h9QI0y8j"
      },
      "source": [
        "1. 荒木雅弘：”フリーソフトでつくる音声認識システム”，森北出版株式会社，2018．\n",
        "2. [声優100人をキレイに話者識別するための機械学習レシピ - Qiita](https://qiita.com/adumaru0828/items/a95de3a0fbfe54f51953)\n",
        "3. [librosa - librosa latest documentation](https://librosa.org/doc/latest/index.html)\n",
        "3. [Welcome to PyCaret - PyCaret Official](https://pycaret.gitbook.io/docs/)\n",
        "4. [声の特性を決めるもの - 国際電気通信基礎技術研究所](https://results.atr.jp/atrj/ATRJ_36/06/main.html)"
      ]
    }
  ],
  "metadata": {
    "colab": {
      "provenance": [],
      "collapsed_sections": [
        "25VCYJjmUBGS",
        "2sxdGRJjTzTU",
        "0gulpd12tHMh",
        "POOzoMhT1zZ2",
        "CZuOyg6nT3Iu",
        "jDfqMH7B_RdT",
        "1Ynmgl3Z_Yju",
        "53pHD6Cq_i_g",
        "CUVaJ9-0K3n1",
        "NDrRbftxUnbP",
        "UE9wmQC7UgAF",
        "B7DrwgvqjB4S"
      ],
      "authorship_tag": "ABX9TyNxl5NtMBxDJVi6ywINxNxe",
      "include_colab_link": true
    },
    "gpuClass": "standard",
    "kernelspec": {
      "display_name": "Python 3",
      "name": "python3"
    },
    "language_info": {
      "name": "python"
    },
    "widgets": {
      "application/vnd.jupyter.widget-state+json": {
        "0fbcc8c8323344f68c5514621b16a3f9": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "IntProgressModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "IntProgressModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "ProgressView",
            "bar_style": "",
            "description": "Processing: ",
            "description_tooltip": null,
            "layout": "IPY_MODEL_8969511da9ad415599e0c7902a0cda2a",
            "max": 69,
            "min": 0,
            "orientation": "horizontal",
            "style": "IPY_MODEL_cf5f6a1e672f4814a2bd2d8aad9e5348",
            "value": 69
          }
        },
        "8969511da9ad415599e0c7902a0cda2a": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "cf5f6a1e672f4814a2bd2d8aad9e5348": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "ProgressStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "ProgressStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "bar_color": null,
            "description_width": ""
          }
        },
        "610467b19ad846f88f4b710f210766a4": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "IntProgressModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "IntProgressModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "ProgressView",
            "bar_style": "",
            "description": "Processing: ",
            "description_tooltip": null,
            "layout": "IPY_MODEL_c8f9271961c14406a4555f393efbe555",
            "max": 7,
            "min": 0,
            "orientation": "horizontal",
            "style": "IPY_MODEL_553005b76fff4f8b9eec8b2dd338a6a0",
            "value": 7
          }
        },
        "c8f9271961c14406a4555f393efbe555": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "553005b76fff4f8b9eec8b2dd338a6a0": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "ProgressStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "ProgressStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "bar_color": null,
            "description_width": ""
          }
        }
      }
    }
  },
  "nbformat": 4,
  "nbformat_minor": 0
}