tan.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "id": "initial_id",
   "metadata": {
    "collapsed": true,
    "ExecuteTime": {
     "end_time": "2024-06-10T10:24:09.442738Z",
     "start_time": "2024-06-10T10:24:09.435618Z"
    }
   },
   "source": [
    "import json\n",
    "\n",
    "import numpy as np\n",
    "from sklearn.metrics import roc_curve\n",
    "from sklearn.metrics import roc_auc_score\n",
    "from naiveBayes import *\n",
    "from sklearn.model_selection import train_test_split"
   ],
   "outputs": [],
   "execution_count": 32
  },
  {
   "metadata": {
    "ExecuteTime": {
     "end_time": "2024-06-10T10:25:18.201717Z",
     "start_time": "2024-06-10T10:25:18.135288Z"
    }
   },
   "cell_type": "code",
   "source": [
    "# 读取数据\n",
    "with open('tan/docs.txt', 'r', encoding='utf-8') as file:\n",
    "    content = file.read()\n",
    "    docs = eval(content)\n",
    "\n",
    "with open('tan/label.txt', 'r', encoding='utf-8') as file:\n",
    "    label = json.load(file)\n",
    "    label = np.array(label)\n",
    "\n",
    "with open('tan/pred.txt', 'r', encoding='utf-8') as file:\n",
    "    content = file.read()\n",
    "    content = content.strip('[]').strip()\n",
    "    tan_pred = list(map(int, content.split()))\n",
    "\n",
    "# 打开文件并读取\n",
    "with open('tan/prob.txt', 'r', encoding='utf-8') as file:\n",
    "    content = file.read()\n",
    "    # 去掉开头和结尾的方括号，以及每行的开头和结尾的方括号\n",
    "    content = content.strip()[1:-1].strip()\n",
    "    # 分割每一行\n",
    "    rows = content.split('\\n')\n",
    "    # 解析每一行\n",
    "    tan_probs = [list(map(float, row.strip()[1:-1].split())) for row in rows]"
   ],
   "id": "cb896382ecb63271",
   "outputs": [],
   "execution_count": 45
  },
  {
   "metadata": {
    "ExecuteTime": {
     "end_time": "2024-06-10T10:25:20.897492Z",
     "start_time": "2024-06-10T10:25:19.188258Z"
    }
   },
   "cell_type": "code",
   "source": [
    "vocabList = createVocabList(docs)\n",
    "# 构建词向量矩阵\n",
    "trainMat = []\n",
    "for inputSet in tqdm(docs, desc='构建词向量矩阵'):\n",
    "    trainMat.append(setOfWords2Vec(vocabList, inputSet))\n",
    "\n",
    "trainMat= np.array(trainMat)\n",
    "X_train, X_test, y_train, y_test = train_test_split(trainMat, label, test_size=0.2, random_state=1)\n"
   ],
   "id": "3087acd2fb0b3f8f",
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "创建词表: 100%|██████████| 1000/1000 [00:00<00:00, 999357.64it/s]\n",
      "构建词向量矩阵: 100%|██████████| 1000/1000 [00:01<00:00, 830.70it/s]\n"
     ]
    }
   ],
   "execution_count": 46
  },
  {
   "metadata": {
    "ExecuteTime": {
     "end_time": "2024-06-10T10:25:20.913606Z",
     "start_time": "2024-06-10T10:25:20.901540Z"
    }
   },
   "cell_type": "code",
   "source": [
    "# tan\n",
    "tan_probs=np.array(tan_probs)[:, 1]\n",
    "fpr_tan, tpr_tan, _ = roc_curve(y_test, tan_probs)\n",
    "roc_auc_tan = roc_auc_score(y_test, tan_probs)"
   ],
   "id": "53c8d7109a0ab2b7",
   "outputs": [],
   "execution_count": 47
  },
  {
   "metadata": {
    "ExecuteTime": {
     "end_time": "2024-06-10T10:25:21.642947Z",
     "start_time": "2024-06-10T10:25:21.635715Z"
    }
   },
   "cell_type": "code",
   "source": [
    "from sklearn.naive_bayes import MultinomialNB\n",
    "from sklearn.svm import SVC\n",
    "\n",
    "from naiveBayes import *\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "from sklearn.metrics import roc_curve, roc_auc_score\n",
    "from sklearn.model_selection import train_test_split\n",
    "import naiveBayesv1 as v1"
   ],
   "id": "313454a21521ce70",
   "outputs": [],
   "execution_count": 48
  },
  {
   "metadata": {
    "ExecuteTime": {
     "end_time": "2024-06-10T10:25:26.830488Z",
     "start_time": "2024-06-10T10:25:22.481161Z"
    }
   },
   "cell_type": "code",
   "source": [
    "# SVM\n",
    "svm = SVC(kernel='linear',probability=True)\n",
    "svm.fit(X_train, y_train)\n",
    "\n",
    "svm_pred = svm.predict(X_test)\n",
    "svm_probs = svm.predict_proba(X_test)[:, 1] "
   ],
   "id": "2a291a5c372613f1",
   "outputs": [],
   "execution_count": 49
  },
  {
   "metadata": {
    "ExecuteTime": {
     "end_time": "2024-06-10T10:25:26.848646Z",
     "start_time": "2024-06-10T10:25:26.834542Z"
    }
   },
   "cell_type": "code",
   "source": [
    "fpr_svm, tpr_svm, _ = roc_curve(y_test, svm_probs)\n",
    "roc_auc_svm = roc_auc_score(y_test, svm_probs)"
   ],
   "id": "f7532cad291526e0",
   "outputs": [],
   "execution_count": 50
  },
  {
   "metadata": {
    "ExecuteTime": {
     "end_time": "2024-06-10T10:25:27.797675Z",
     "start_time": "2024-06-10T10:25:27.735242Z"
    }
   },
   "cell_type": "code",
   "source": [
    "# 朴素贝叶斯\n",
    "nbmodel = MultinomialNB()\n",
    "nbmodel.fit(X_train, y_train)\n",
    "\n",
    "nb_pred = nbmodel.predict(X_test)\n",
    "nb_probs = nbmodel.predict_proba(X_test)[:, 1] "
   ],
   "id": "d30aab4c0873f25a",
   "outputs": [],
   "execution_count": 51
  },
  {
   "metadata": {
    "ExecuteTime": {
     "end_time": "2024-06-10T10:25:28.620702Z",
     "start_time": "2024-06-10T10:25:28.606451Z"
    }
   },
   "cell_type": "code",
   "source": [
    "fpr_nb, tpr_nb, _ = roc_curve(y_test, nb_probs)\n",
    "roc_auc_nb = roc_auc_score(y_test, nb_probs)"
   ],
   "id": "44fc92dad006c9ee",
   "outputs": [],
   "execution_count": 52
  },
  {
   "metadata": {
    "ExecuteTime": {
     "end_time": "2024-06-10T10:25:29.688150Z",
     "start_time": "2024-06-10T10:25:29.655159Z"
    }
   },
   "cell_type": "code",
   "source": [
    "class LinearSVM:\n",
    "    def __init__(self, learning_rate=0.001, lambda_param=0.01, n_iters=1000):\n",
    "        self.lr = learning_rate\n",
    "        self.lambda_param = lambda_param\n",
    "        self.n_iters = n_iters\n",
    "        self.w = None\n",
    "        self.b = None\n",
    "\n",
    "    def fit(self, X, y):\n",
    "        n_samples, n_features = X.shape\n",
    "        self.w = np.zeros(n_features)\n",
    "        self.b = 0\n",
    "\n",
    "        for _ in tqdm(range(self.n_iters), desc='训练SVM'):\n",
    "            for idx, x_i in enumerate(X):\n",
    "                condition = y[idx] * (np.dot(x_i, self.w) - self.b) >= 1\n",
    "                if condition:\n",
    "                    self.w -= self.lr * (2 * self.lambda_param * self.w)\n",
    "                else:\n",
    "                    self.w -= self.lr * (2 * self.lambda_param * self.w - np.dot(x_i, y[idx]))\n",
    "                    self.b -= self.lr * y[idx]\n",
    "\n",
    "    def predict(self, X):\n",
    "        return np.sign(np.dot(X, self.w) - self.b)\n",
    "\n",
    "    def decision_function(self, X):\n",
    "        return np.dot(X, self.w) - self.b"
   ],
   "id": "6b8ffaee70060147",
   "outputs": [],
   "execution_count": 53
  },
  {
   "metadata": {
    "ExecuteTime": {
     "end_time": "2024-06-10T10:25:30.258260Z",
     "start_time": "2024-06-10T10:25:30.251434Z"
    }
   },
   "cell_type": "code",
   "source": [
    "def euclidean_distance(x1, x2):\n",
    "    return np.sqrt(np.sum((x1 - x2) ** 2))"
   ],
   "id": "a83bb8935ce7d9d3",
   "outputs": [],
   "execution_count": 54
  },
  {
   "metadata": {
    "ExecuteTime": {
     "end_time": "2024-06-10T10:25:30.720979Z",
     "start_time": "2024-06-10T10:25:30.707039Z"
    }
   },
   "cell_type": "code",
   "source": [
    "def tsvm_nb_algorithm(X, y):\n",
    "    # 初始化距离矩阵\n",
    "    n_samples = len(X)\n",
    "    distance_matrix = np.full((n_samples, n_samples), np.inf)\n",
    "\n",
    "    for i in tqdm(range(n_samples), desc='计算距离矩阵'):\n",
    "        for j in range(n_samples):\n",
    "            if i != j:\n",
    "                distance_matrix[i, j] = euclidean_distance(X[i], X[j])\n",
    "\n",
    "    # 初始化每个点的最近邻和最短距离\n",
    "    nearest_neighbors = np.zeros(n_samples, dtype=int)\n",
    "    min_distances = np.full(n_samples, np.inf)\n",
    "\n",
    "    for i in tqdm(range(n_samples), desc='选择最近邻'):\n",
    "        for j in range(n_samples):\n",
    "            if distance_matrix[i, j] < min_distances[i]:\n",
    "                min_distances[i] = distance_matrix[i, j]\n",
    "                nearest_neighbors[i] = j\n",
    "\n",
    "    # 初始化标志矩阵\n",
    "    flags = np.ones(n_samples)\n",
    "\n",
    "    for i in tqdm(range(n_samples), desc='计算标志矩阵'):\n",
    "        neighbor_idx = nearest_neighbors[i]\n",
    "        if y[i] != y[neighbor_idx]:\n",
    "            flags[i] = -1\n",
    "        else:\n",
    "            flags[i] = 1\n",
    "\n",
    "    # 收集需要删除的索引\n",
    "    delete_indices = set()\n",
    "\n",
    "    for i in tqdm(range(n_samples), desc='修剪样本集'):\n",
    "        neighbor_idx = nearest_neighbors[i]\n",
    "        if flags[i] == -1:\n",
    "            # 选择删除点，优先删除距离较远的点\n",
    "            if min_distances[i] < min_distances[neighbor_idx]:\n",
    "                delete_indices.add(i)\n",
    "            else:\n",
    "                delete_indices.add(neighbor_idx)\n",
    "\n",
    "    # 统一删除收集到的索引\n",
    "    delete_indices = sorted(delete_indices, reverse=True)\n",
    "    for idx in delete_indices:\n",
    "        X = np.delete(X, idx, axis=0)\n",
    "        y = np.delete(y, idx, axis=0)\n",
    "\n",
    "    # 再次用NB算法训练\n",
    "    nb = NaiveBayes()\n",
    "    nb.fit(X, y)\n",
    "\n",
    "    return nb"
   ],
   "id": "d750bde512543637",
   "outputs": [],
   "execution_count": 55
  },
  {
   "metadata": {
    "ExecuteTime": {
     "end_time": "2024-06-10T10:26:22.261923Z",
     "start_time": "2024-06-10T10:25:31.739451Z"
    }
   },
   "cell_type": "code",
   "source": [
    "# 初始训练\n",
    "nb_initial = NaiveBayes()\n",
    "nb_initial.fit(X_train, y_train)\n",
    "initial_predictions = nb_initial.predict(X_train)\n",
    "\n",
    "# 构建最优分类超平面\n",
    "svm = LinearSVM()\n",
    "svm.fit(X_train, initial_predictions)\n",
    "\n",
    "distances = svm.decision_function(X_train)\n",
    "threshold = 0.2\n",
    "selected_samples = np.abs(distances) > threshold\n",
    "\n",
    "X_optimized = X_train[selected_samples]\n",
    "y_optimized = y_train[selected_samples]\n",
    "\n",
    "nb = tsvm_nb_algorithm(X_optimized, y_optimized)\n",
    "y_pred = nb.predict(X_test)\n",
    "y_probs = nb.predict_proba(X_test)[:, 1]  # 选择概率中的正类概率"
   ],
   "id": "feb30b7185bf2106",
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "训练SVM: 100%|██████████| 1000/1000 [00:31<00:00, 31.77it/s]\n",
      "计算距离矩阵: 100%|██████████| 800/800 [00:18<00:00, 43.12it/s]\n",
      "选择最近邻: 100%|██████████| 800/800 [00:00<00:00, 2354.54it/s]\n",
      "计算标志矩阵: 100%|██████████| 800/800 [00:00<00:00, 403783.78it/s]\n",
      "修剪样本集: 100%|██████████| 800/800 [00:00<00:00, 727704.01it/s]\n"
     ]
    }
   ],
   "execution_count": 56
  },
  {
   "metadata": {
    "ExecuteTime": {
     "end_time": "2024-06-10T10:26:31.179931Z",
     "start_time": "2024-06-10T10:26:31.164350Z"
    }
   },
   "cell_type": "code",
   "source": [
    "fpr_TSVM, tpr_TSVM, _ = roc_curve(y_test, y_probs)\n",
    "roc_auc_TSVM = roc_auc_score(y_test, y_probs)"
   ],
   "id": "79daec6fddf6439b",
   "outputs": [],
   "execution_count": 57
  },
  {
   "metadata": {
    "ExecuteTime": {
     "end_time": "2024-06-10T10:26:32.825927Z",
     "start_time": "2024-06-10T10:26:32.251904Z"
    }
   },
   "cell_type": "code",
   "source": [
    "# 绘制ROC曲线\n",
    "plt.figure()\n",
    "plt.plot(fpr_TSVM, tpr_TSVM, color='darkorange', lw=2, label='TSVM-NB (AUC = %0.2f)' % roc_auc_TSVM)\n",
    "plt.plot(fpr_svm, tpr_svm, color='green', lw=2, label='SVM (AUC = %0.2f)' % roc_auc_svm)\n",
    "plt.plot(fpr_nb, tpr_nb, color='blue', lw=2, label='NaiveBayes (AUC = %0.2f)' % roc_auc_nb)\n",
    "plt.plot(fpr_tan, tpr_tan, color='red', lw=2, label='TAN (AUC = %0.2f)' % roc_auc_tan)\n",
    "\n",
    "# 绘制对角线\n",
    "plt.plot([0, 1], [0, 1], color='navy', lw=2, linestyle='--')\n",
    "\n",
    "# 设置图形的范围、标签和标题\n",
    "plt.xlim([0.0, 1.0])\n",
    "plt.ylim([0.0, 1.05])\n",
    "plt.xlabel('False Positive Rate')\n",
    "plt.ylabel('True Positive Rate')\n",
    "plt.title('Receiver Operating Characteristic (ROC) Curve')\n",
    "plt.legend(loc=\"lower right\")\n",
    "\n",
    "# 显示图形\n",
    "plt.show()"
   ],
   "id": "a76b53951b4c1349",
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ],
      "image/png": "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"
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "execution_count": 58
  },
  {
   "metadata": {
    "ExecuteTime": {
     "end_time": "2024-06-10T10:26:37.888604Z",
     "start_time": "2024-06-10T10:26:37.861582Z"
    }
   },
   "cell_type": "code",
   "source": [
    "from sklearn.metrics import matthews_corrcoef\n",
    "\n",
    "mcc = matthews_corrcoef(y_test, y_pred)\n",
    "mcc_svm=matthews_corrcoef(y_test,svm_pred)\n",
    "mcc_nb=matthews_corrcoef(y_test,nb_pred)\n",
    "mcc_tan=matthews_corrcoef(y_test,tan_pred)\n",
    "print(mcc)\n",
    "print(mcc_svm)\n",
    "print(mcc_nb)\n",
    "print(mcc_tan)"
   ],
   "id": "be7020e4dd0712f7",
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.8845336935740052\n",
      "0.8666053651488208\n",
      "0.8845336935740052\n",
      "0.929539499620758\n"
     ]
    }
   ],
   "execution_count": 59
  },
  {
   "metadata": {},
   "cell_type": "code",
   "execution_count": null,
   "source": "",
   "id": "1a8c4da17fd0067b",
   "outputs": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 2
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython2",
   "version": "2.7.6"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}