-
Notifications
You must be signed in to change notification settings - Fork 0
/
main.py
181 lines (137 loc) · 5.61 KB
/
main.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
import pandas as pd # 데이터 전처리
import numpy as np # 데이터 전처리
from pandas import DataFrame #데이터 전처리
import sys
import math
import preprocess as pp
from data_explanation import *
from keras.models import Sequential
from keras.layers import Dense, Conv1D, MaxPooling1D, Flatten, LSTM, Dropout
def build_model():
#### LSTM
model = Sequential()
model.add(LSTM(100, return_sequences=False, input_shape=(x_shape, 1)))
model.add(Dropout(0.25))
model.add(Dense(y_shape, activation='linear'))
# ##### Conv1D
# model = Sequential()
# model.add(Conv1D(128, 2, activation='relu', input_shape=(x_shape, 1)))
# model.add(MaxPooling1D(2))
# model.add(Conv1D(64, 2, activation='relu'))
# model.add(MaxPooling1D(2))
# model.add(Flatten())
# model.add(Dense(y_shape, activation='linear'))
return model
from sklearn.preprocessing import MinMaxScaler
if __name__ == '__main__':
x_shape = 24*5
y_shape = 24 * 1
underfit_key = []
data_shape = []
weather = pd.read_csv("input/incheon_weather.csv", encoding="euc-kr")
test = pd.read_csv('input/test.csv')
test = pp.overloaded_na_handle(test)
test.to_csv("input/nan_pre.csv", index=False)
submission = pd.read_csv("input/submission_1002.csv")
# # 전처리
test = pd.read_csv('input/nan_pre.csv')
test = pp.drop_dummy(test)
test = pp.fbfill_nan(test)
test = pp.drop_notfullday(test)
try:
test.to_csv("input/test_result.csv", index=False)
except:
print("Save Failed. Permission Denied \n\n")
pass
agg={}
# index = weather.loc[weather.iloc[:, 1] == ].index
weather['Time']=pd.to_datetime(weather['일시'])
test['Time']=pd.to_datetime(test['Time'])
for i in range(1, len(test.columns) ):
key = test.columns[i]
data = test.iloc[:, i]
idx = data.index
data = pp.drop_null(data)
data = data.values
data = data.astype('float32')
data = np.reshape(data, (-1, 1))
scaler = MinMaxScaler(feature_range=(0, 1))
# data = scaler.fit_transform(data)
data = data.reshape(-1)
print(weather.loc[weather.iloc[1, 1]])
start_time = test.iloc[idx[0], 0]
print(start_time)
weather = weather[weather.iloc[: : , :]]
degree, __ = pp.weather_preprocess(weather)
scaler_degree = MinMaxScaler()
degree = degree.reshape(-1, 1)
# degree = scaler_degree.fit_transform(degree)
degree = degree.reshape(-1)
pp.split2(data, degree, x_shape, y_shape, gap=1, debug=True)
sys.exit()
# data_shape.append(x.shape[0])
# x = x.reshape(x.shape[0], x.shape[1], -1 )
# pred_data = pred_data.reshape(pred_data.shape[0], pred_data.shape[1], -1 )
# y=y.squeeze()
# # print(x.shape)
# # print(y.shape)
# # print(pred_data.shape)
# from sklearn.utils import shuffle
# x, y = shuffle(x, y, random_state = 30)
# from sklearn.model_selection import KFold
# kf = KFold(n_splits=15)
# for train_index, test_index in kf.split(x):
# x_train, x_test = x[train_index], x[test_index]
# y_train, y_test = y[train_index], y[test_index]
# # test_size = 50
# # x_train = x[:-test_size, :, :]
# # y_train = y[:-test_size, :]
# # x_test = x[-test_size : , :, :]
# # y_test = y[-test_size : , :]
# print(x.shape)
# print(x_train.shape)
# print(x_test.shape)
# from keras.callbacks import EarlyStopping
# model = build_model()
# model.compile(optimizer='adam', loss='mse')
# early_stopping = EarlyStopping(monitor='val_loss', patience=5, mode='auto')
# model.fit(x_train, y_train, batch_size=512, epochs=2, callbacks=[early_stopping], validation_data=(x_test, y_test), verbose= 0)
# loss = model.evaluate(x_test, y_test, verbose= 0)
# print(loss)
# if loss > 1.0:
# print(loss)
# underfit_key.append(key)
# a = pd.DataFrame()
# pred = model.predict(pred_data)
# pred = scaler.inverse_transform(pred)
# for j in range(24):
# if pred[0][j] < 0:
# print("found value less than 0")
# sys.exit()
# a['X2018_7_1_'+str(j+1)+'h'] = [pred[0][j]]
# for j in range(10):
# a['X2018_7_'+str(j+1)+'_d'] = 0. # column명을 submission 형태에 맞게 지정합니다.
# # 월별 예측
# # 일별로 예측하여 7월 ~ 11월의 일 수에 맞게 나누어 합산합니다.
# a['X2018_7_m'] = [0.] # 7월
# a['X2018_8_m'] = [0.] # 8월
# a['X2018_9_m'] = [0.] # 9월
# a['X2018_10_m'] = [0.] # 10월
# a['X2018_11_m'] = [0.] # 11월
# a['meter_id'] = key
# agg[key] = a[submission.columns.tolist()]
# print(key)
# print("current_index: ", i)
# del model
# del data
# del x, y
# print('\n\n\n')
# data_summary(data_shape)
# print('---- Modeling Done ----')
# print("Model with loss higher than 1.0\n", underfit_key)
# print("count", len(underfit_key))
# output1 = pd.concat(agg, ignore_index=False)
# output2 = output1.reset_index().drop(['level_0','level_1'], axis=1)
# output2['id'] = output2['meter_id'].str.replace('X','').astype(int)
# output2 = output2.sort_values(by='id', ascending=True).drop(['id'], axis=1).reset_index(drop=True)
# output2.to_csv('prediction.csv', index=False)