-
Notifications
You must be signed in to change notification settings - Fork 20
/
Copy pathdata_transformation.py
267 lines (227 loc) · 9.25 KB
/
data_transformation.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
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
# -*- coding:utf-8 -*-
import pandas as pd
import numpy as np
from sklearn.model_selection import train_test_split
from sklearn.metrics import accuracy_score
from sklearn.decomposition import PCA
def split_data(data_x, data_y, sample0_test_proportion, sample1_test_proportion, sample2_test_proportion, random=1):
"""
用随机的方法，按照指定的各类测试集样本比例生成训练集和测试集
:param data_x: 样本, np.array
:param data_y: 标签,[]
:param sample0_test_proportion: 0类样本测试集比例
:param sample1_test_proportion: 1类样本测试集比例
:param sample2_test_proportion: 2类样本测试集比例
:return:
"""
data0_x, data1_x, data2_x = [], [], []
data0_y, data1_y, data2_y = [], [], []
for x, y in zip(data_x, data_y):
if y == 0:
data0_x.append(x)
data0_y.append(y)
elif y == 1:
data1_x.append(x)
data1_y.append(y)
else:
data2_x.append(x)
data2_y.append(y)
if random:
x_train_0, x_test_0, y_train_0, y_test_0 = train_test_split(data0_x, data0_y, test_size=sample0_test_proportion,
random_state=8)
x_train_1, x_test_1, y_train_1, y_test_1 = train_test_split(data1_x, data1_y, test_size=sample1_test_proportion,
random_state=8)
x_train_2, x_test_2, y_train_2, y_test_2 = train_test_split(data2_x, data2_y, test_size=sample2_test_proportion,
random_state=8)
x_train = np.concatenate((x_train_0, x_train_1, x_train_2), axis=0)
y_train = np.concatenate((y_train_0, y_train_1, y_train_2), axis=0)
if x_test_0:
x_test = np.concatenate((x_test_0, x_test_1, x_test_2), axis=0)
y_test = np.concatenate((y_test_0, y_test_1, y_test_2), axis=0)
else:
x_test = []
y_test = []
else:
x_train = np.concatenate((data0_x, data1_x, data2_x), axis=0)
x_test = []
y_train = np.concatenate((data0_y, data1_y, data2_y), axis=0)
y_test = []
return x_train, x_test, y_train, y_test
def generate_y(times, t1, t2):
"""
根据时间序列生成标签
:param times:
:return:[]
"""
data_y = []
for time in times:
if time <= t1:
data_y.append(0)
elif time <= t2:
data_y.append(1)
else:
data_y.append(2)
return data_y
def accuracy(test_labels, predict_labels):
line1, line2 = 0, 0
for i in test_labels:
if i == 0:
line1 += 1
elif i == 1:
line2 += 1
line2 += line1
accuracy_score0 = accuracy_score(test_labels[0:line1], predict_labels[0:line1])
accuracy_score1 = accuracy_score(test_labels[line1:line2], predict_labels[line1:line2])
accuracy_score2 = accuracy_score(test_labels[line2:], predict_labels[line2:])
return accuracy_score0, accuracy_score1, accuracy_score2
def weight_of_prediction(y, weights):
"""
根据多个信号的预测结果，结合权重输出最终预测结果
:param y: [y1,y2,...,yn] 0,1,2
:param weights: [w1,w2,...,wm]
:return:
"""
if set(y) == {0} or set(y) == {1} or set(y) == {2}:
return y[0]
elif set(y) == {0, 1}:
return np.argmax(np.array(weights[:2]) * np.bincount(y))
else:
return np.argmax(np.array(weights)*np.bincount(y))
def prediction_use_prob(probas, num_samples):
"""
根据概率计算分类（加权）
:param prob:
:param num_samples: 投票的信号个数
:return:
"""
predict_y = []
# w = [1.04, 1., 1.03]
# for i in range(probas.shape[0]):
# probas[i] = (probas[i]*w)/sum(probas[i]*w)
probas = probas*[1., 1., 1.]
for i in range(0, len(probas), num_samples):
p = list(sum(probas[i:i + num_samples, :]))
predict_y.append(p.index(max(p)))
return predict_y
def one_tensor_data(data_file, time1, time2, sensor_index):
"""
:param data_file:
:param time1:
:param time2:
:param sensor_index:
:return: data_x:pd.DataFrame data_y:[]
"""
with open(data_file, 'r') as fl:
data = pd.read_csv(fl)
data_1 = data.loc[data['通道'] == sensor_index, :]
times = data_1.iloc[:, 1]
data_x = data_1.drop(['信号强度', '初始频率', '绝对能量', '中心频率', '峰频'], axis=1).iloc[:, 2:]
data_y = generate_y(times, time1, time2)
return data_x, data_y
def four_tensor_data(data_file, time1, time2, type=0):
"""
输出四个通道的数据
:param data_file:
:return:data_x:pd.DataFrame, data_y:[]
"""
with open(data_file, 'r') as fl:
data = pd.read_csv(fl)
if not type:
times = data.iloc[:, 1]
data_x = data.iloc[:, 2:]
data_y = generate_y(times, time1, time2)
return data_x, data_y
times = data.iloc[:, 1]
data_x = data.drop(['信号强度', '初始频率', '绝对能量', '中心频率', '峰频'], axis=1).iloc[:, 2:]
data_y = generate_y(times, time1, time2)
return data_x, data_y
def four_tensor_frequency(frequency_data_file, time1, time2):
with open(frequency_data_file, 'r') as fl:
f_data = pd.read_csv(fl)
times = f_data['TIME OF TEST']
data_x = f_data.iloc[:, 2:]
data_y = generate_y(times, time1, time2)
return data_x, data_y
def four_tensor_data_added_frequency(data_file, frequency_data_file, time1, time2, pca_n=None):
with open(data_file, 'r') as fl:
data = pd.read_csv(fl)
times = data.iloc[:, 0]
data_x = data.drop(['信号强度', '初始频率', '绝对能量', '中心频率', '峰频'], axis=1).iloc[:, 2:]
data_y = generate_y(times, time1, time2)
with open(frequency_data_file, 'r') as fl:
f_data = pd.read_csv(fl)
f_data_x = f_data.iloc[:, 2:]
if pca_n:
f_data_x = pd.DataFrame(data=PCA(n_components=pca_n).fit_transform(f_data_x))
data_x = pd.concat([data_x, f_data_x], axis=1)
return data_x, data_y
def add_frequency(data_file, frequency_data_file, sensor_index, time1, time2):
with open(frequency_data_file, 'r') as fl:
f_data = pd.read_csv(fl)
# CHANNEL NUMBER, TIME OF TEST, 0, 1, 2 ... 198, 199
f_data_x = f_data.loc[f_data['CHANNEL NUMBER'] == sensor_index, :].iloc[:, 2:]
times = f_data['TIME OF TEST']
with open(data_file, 'r') as fl:
data = pd.read_csv(fl)
data_1 = data.loc[data['通道'] == sensor_index, :].iloc[:, 2:-1]
data_x = data_1.drop(['信号强度', '初始频率', '绝对能量', '中心频率', '峰频'], axis=1)
data_x = pd.concat([data_x, f_data_x], axis=1)
data_y = generate_y(times, time1, time2)
return data_x, data_y
# ----------------------for cnn model------------------------
def fft_generate_y(times, t1, t2):
"""
根据时间序列生成标签
:param times:
:return:
"""
data_y = []
for time in times:
if time <= t1:
data_y.append([1, 0, 0])
elif time <= t2:
data_y.append([0, 1, 0])
else:
data_y.append([0, 0, 1])
return np.array(data_y)
def fft_four_tensor_frequency(frequency_data_file, time1, time2):
with open(frequency_data_file, 'r') as fl:
f_data = pd.read_csv(fl)
times = f_data['TIME OF TEST']
data_x = f_data.iloc[:, 2:]
data_y = fft_generate_y(times, time1, time2)
return data_x, data_y
def fft_split_data(data_x, data_y, sample0_test_proportion, sample1_test_proportion, sample2_test_proportion):
"""
用随机的方法，按照指定的各类测试集样本比例生成训练集和测试集
:param data_x: 样本, np.array
:param data_y: 标签, np.array
:param sample0_test_proportion: 0类样本测试集比例
:param sample1_test_proportion: 1类样本测试集比例
:param sample2_test_proportion: 2类样本测试集比例
:return:
"""
data0_x, data1_x, data2_x = [], [], []
data0_y, data1_y, data2_y = [], [], []
for x, y in zip(data_x, data_y):
if y[0] == 1:
data0_x.append(x)
data0_y.append(y)
elif y[1] == 1:
data1_x.append(x)
data1_y.append(y)
else:
data2_x.append(x)
data2_y.append(y)
x_train_0, x_test_0, y_train_0, y_test_0 = train_test_split(data0_x, data0_y, test_size=sample0_test_proportion,
random_state=8)
x_train_1, x_test_1, y_train_1, y_test_1 = train_test_split(data1_x, data1_y, test_size=sample1_test_proportion,
random_state=8)
x_train_2, x_test_2, y_train_2, y_test_2 = train_test_split(data2_x, data2_y, test_size=sample2_test_proportion,
random_state=8)
x_train = np.concatenate((x_train_0, x_train_1, x_train_2), axis=0)
x_test = np.concatenate((x_test_0, x_test_1, x_test_2), axis=0)
y_train = np.concatenate((y_train_0, y_train_1, y_train_2), axis=0)
y_test = np.concatenate((y_test_0, y_test_1, y_test_2), axis=0)
return x_train, x_test, y_train, y_test
# ----------------------for cnn model------------------------