-
Notifications
You must be signed in to change notification settings - Fork 0
/
main.py
361 lines (314 loc) · 15.6 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
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
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
# -*- coding: utf-8 -*-
import pandas as pd
import numpy as np
import draw_figure as df
from StatsToolKits import StatsToolKits
from toolkits import ToolKits
from FileHelper import FileHelper
from os import listdir
from os.path import isfile, join
from enums import BPTypes, BHSTypes
from sklearn.ensemble import RandomForestRegressor
from sklearn.cross_validation import train_test_split, cross_val_score, ShuffleSplit, KFold
# 类名 与 属性名 要采用 驼峰表达式
# 方法名 与 局部变量名 要采用 下划线表达式
# 使用‘\’来换行 但在[]/()/{}中无需这样使用
class RegressionAlgorithm:
cols = ['hr_miu', 'hr_delta', 'hr_iqr', 'hr_skew',
'pwtt_mean', 'PH', 'PRT', 'PWA', 'RBAr', 'SLP1', 'SLP2',
'RBW10', 'RBW25', 'RBW33', 'RBW50', 'RBW66', 'RBW75',
'DBW10', 'DBW25', 'DBW33', 'DBW50', 'DBW66', 'DBW75',
'KVAL', 'AmBE', 'DfAmBE',
'kte_miu', 'kte_delta', 'kte_iqr', 'kte_skew',
'h_miu', 'h_delta', 'h_iqr', 'h_skew',
'loge_ar_1', 'loge_ar_2', 'loge_ar_3', 'loge_ar_4', 'loge_ar_5',
'loge_delta', 'loge_iqr', 'ppg_fed_ar_1', 'ppg_fed_ar_2',
'ppg_fed_ar_3', 'ppg_fed_ar_4', 'ppg_fed_ar_5']
colsRes = ['sbps', 'dbps'] # 'sbps' , 'dbps'
colsResTypes = [BPTypes.SBP, BPTypes.DBP]
minFullSetSize = 30 # 只有全集大于等于30时才进行训练
def __init__(self):
self.x_train = []
self.y_train = []
self.x_test = []
self.y_test = []
self.testResults = []
self.rf = RandomForestRegressor(n_estimators=100, max_depth=4, warm_start=True)
self.dt = df.DrawToolkit() # plot assistant
self.type = BPTypes.SBP
def train(self):
train_arr = self.x_train
train_res = self.y_train[:, self.colsResTypes.index(self.type)]
self.rf.fit(train_arr, train_res)
def train_mbp(self):
self.rf.fit(self.x_train, self.y_train)
def test(self):
test_arr = self.x_test
self.testResults = self.rf.predict(test_arr)
return self.testResults
def get_test_set_original_results(self):
return list(self.y_test[:, self.colsResTypes.index(self.type)])
def show_full_set_result(self, title):
tmp_plt = self.dt.generate_scatter_plt(self.get_test_set_original_results(), list(self.testResults),
"Measured " + BPTypes.get_type_name(self.type) +
"(mmHg)", "Estimated " + BPTypes.get_type_name(self.type) + "(mmHg)",
title)
tmp_plt.show()
def show_mbp_full_set_result(self, title, mbp_local):
tmp_plt = self.dt.generate_scatter_plt(mbp_local, list(self.testResults),
"Measured " + BPTypes.get_type_name(self.type) +
"(mmHg)", "Estimated " + BPTypes.get_type_name(self.type) + "(mmHg)",
title)
tmp_plt.show()
def save_predict_result(self, sig_name, title, plt_root_path):
tmp_plt = self.dt.generate_scatter_plt(list(self.y_test[:, self.colsResTypes.index(self.type)]), list(self.testResults),
"Measured " + BPTypes.get_type_name(self.type) +
"(mmHg)", "Estimated " + BPTypes.get_type_name(self.type) + "(mmHg)",
sig_name + ": " + title)
tmp_plt.savefig(plt_root_path + '_' + BPTypes.get_type_name(self.type) + '_' + sig_name + '.png')
def get_result_bhs_type(self):
small_than_5, small_than_10, small_than_15 = [0] * 3
for val in list(map(lambda x: x[0]-x[1],
zip(self.testResults, self.y_test[:, self.colsResTypes.index(self.type)]))):
val = abs(val)
if val <= 15:
small_than_15 += 1
if val <= 10:
small_than_10 += 1
if val <= 5:
small_than_5 += 1
return ToolKits.which_bhs_standard(100.0 * small_than_5/len(self.testResults),
100.0 * small_than_10/len(self.testResults),
100.0 * small_than_15/len(self.testResults))
def read_file(self, data_file_name):
full_set = pd.read_csv(data_file_name)
valid, arrs, indexes = self.pick_data(full_set.as_matrix(self.cols))
return valid, arrs, full_set.as_matrix(self.colsRes) if valid else []
def split_sets(self, src_set, res_set):
self.x_train, self.x_test, self.y_train, self.y_test = train_test_split(src_set, res_set)
def set_data_type(self, bp_type):
self.type = bp_type
def reset_model(self):
self.rf = RandomForestRegressor(n_estimators=100, max_depth=4, warm_start=True)
def alter_type(self):
self.type = self.colsResTypes[0] if self.type == self.colsResTypes[1] else self.colsResTypes[1]
def pick_data(self, data): # find column which contains 0 and eliminate them
is_legal = False
for_return = np.matrix([])
indexes = []
for index_local in range(0, data.shape[0]):
indexes.append(index_local)
if np.where(data[index_local] == 0).size == 0:
if for_return.size == 0:
is_legal = True
for_return = data[index_local]
else:
for_return = np.concatenate((for_return, data[index_local]), axis=0)
return is_legal, for_return, indexes
def get_title(rf_local):
# when you are asked to "shadow name", you should change the name.
x = rf_local.get_test_set_original_results()
y = rf_local.testResults
type_str = BPTypes.get_type_name(rf_local.type)
stats_toolkit_obj = StatsToolKits(x, y)
pearson_r, p_val = stats_toolkit_obj.get_pearson_corr()
mic = stats_toolkit_obj.get_mic()
return type_str + ' ' + 'Regression Results\n' + \
'pearson regression: ' + '%.2f' % pearson_r + ' p value: ' + '%.2e' % p_val + '\n'\
'maximal information coefficient: ' + '%.2f' % mic
def disp_stats_paras(rf_local):
stats_toolkit_obj = StatsToolKits([], [])
print("testset size:" + "%d" % (len(rf_local.get_test_set_original_results()) / 0.2))
print("testset range:" + "%d" % stats_toolkit_obj.get_range(rf_local.get_test_set_original_results()))
print("standard derivation:" + "%.2f" % stats_toolkit_obj.get_std(rf_local.get_test_set_original_results()))
def rank_features(x, y, feature_names):
rf_local = RandomForestRegressor(n_estimators=20, max_depth=4)
scores = []
for i in range(x.shape[1]):
score = cross_val_score(rf_local, x[:, i:i+1], y, scoring="r2", cv=ShuffleSplit(len(x), 3, .3))
scores.append((round(np.mean(score), 3), feature_names[i]))
return sorted(scores, reverse=True)
def disp_map(map_val):
for val in map_val:
print(str(val[0]) + '\t' + str(val[1]))
def get_mean_bp(sbp_and_dbp_mat): # 获取到mbp # 给两列分别为SBP和DBP的matrix添加第三列（2dbp+sbp）/3
mbp = (2 * sbp_and_dbp_mat[:, 0] + sbp_and_dbp_mat[:, 1]) / 3
return mbp
# return np.concatenate((sbp_and_dbp_mat, mbp), axis=1)
if __name__ == "__main__":
# a = [(0.098, 'RBW10'), (0.095, 'RBW25'), (0.089, 'RBW33'), (0.087, 'kte_delta'), (0.076, 'RBW50'), (0.069, 'pwtt_mean'), (0.067, 'RBW66'), (0.064, 'kte_skew'), (0.058, 'RBW75'), (0.058, 'DBW50'), (0.057, 'SLP2'), (0.057, 'SLP1'), (0.056, 'h_miu'), (0.055, 'kte_iqr'), (0.055, 'PRT'), (0.055, 'DBW33'), (0.053, 'DBW25'), (0.051, 'DBW10'), (0.049, 'DBW66'), (0.042, 'KVAL'), (0.042, 'DBW75'), (0.034, 'hr_delta'), (0.028, 'hr_miu'), (0.028, 'PH'), (0.027, 'ppg_fed_ar_5'), (0.027, 'RBAr'), (0.027, 'AmBE'), (0.026, 'kte_miu'), (0.026, 'PWA'), (0.021, 'DfAmBE'), (0.017, 'hr_skew'), (0.014, 'hr_iqr'), (0.013, 'ppg_fed_ar_4'), (0.009, 'ppg_fed_ar_1'), (0.009, 'h_delta'), (0.008, 'ppg_fed_ar_2'), (0.008, 'loge_delta'), (0.006, 'loge_iqr'), (0.005, 'h_iqr'), (0.002, 'ppg_fed_ar_3'), (0.001, 'h_skew'), (-0.0, 'loge_ar_5'), (-0.0, 'loge_ar_4'), (-0.0, 'loge_ar_3'), (-0.0, 'loge_ar_2'), (0.0, 'loge_ar_1')]
# b = [(0.086, 'kte_delta'), (0.064, 'hr_miu'), (0.055, 'kte_iqr'), (0.054, 'RBW75'), (0.052, 'RBW66'), (0.049, 'RBW50'), (0.045, 'hr_delta'), (0.043, 'RBW33'), (0.037, 'RBW25'), (0.035, 'PWA'), (0.035, 'DBW66'), (0.034, 'kte_miu'), (0.034, 'RBW10'), (0.034, 'DBW50'), (0.033, 'AmBE'), (0.032, 'pwtt_mean'), (0.032, 'DfAmBE'), (0.028, 'RBAr'), (0.028, 'PH'), (0.028, 'DBW75'), (0.026, 'DBW33'), (0.026, 'DBW25'), (0.026, 'DBW10'), (0.024, 'ppg_fed_ar_2'), (0.023, 'ppg_fed_ar_1'), (0.022, 'kte_skew'), (0.022, 'KVAL'), (0.021, 'SLP1'), (0.017, 'hr_skew'), (0.017, 'PRT'), (0.015, 'SLP2'), (0.014, 'hr_iqr'), (0.013, 'ppg_fed_ar_5'), (0.012, 'h_miu'), (0.01, 'ppg_fed_ar_3'), (0.009, 'loge_delta'), (0.009, 'h_iqr'), (0.007, 'loge_iqr'), (0.007, 'h_delta'), (0.003, 'ppg_fed_ar_4'), (0.002, 'loge_ar_1'), (0.002, 'h_skew'), (0.001, 'loge_ar_2'), (0.0, 'loge_ar_5'), (0.0, 'loge_ar_4'), (-0.0, 'loge_ar_3')]
# disp_map(a)
# print('*****************')
# disp_map(b)
# exit()
fh = FileHelper()
all_csv_names = fh.get_all_csv_names()
full_set_arr = []
full_set_res = []
for csv_file_name in all_csv_names:
is_valid, arr, res = fh.read_file(csv_file_name)
if not is_valid:
continue
full_set_arr = np.concatenate((full_set_arr, arr.tolist()), 0) if len(full_set_arr) > 0 else arr.tolist()
res = get_mean_bp(res)
full_set_res = np.concatenate((full_set_res, res.tolist()), 0) if len(full_set_res) > 0 else res.tolist()
# print '************' + fh.colsRes[0] + '***************'
# print rank_features(full_set_arr, full_set_res[:, 0], fh.cols)
# print '************' + fh.colsRes[1] + '***************'
# print rank_features(full_set_arr, full_set_res[:, 1], fh.cols)
kf = KFold(full_set_res.shape[0], 10)
max_sbp_corr = 0
max_dbp_corr = 0
sbp_corrs = []
dbp_corrs = []
max_sbp_train_index = []
max_sbp_test_index = []
max_dbp_train_index = []
max_dbp_test_index = []
index = 0
max_sbp_index = 0
max_dbp_index = 0
max_sbp_index_num = 6
max_dbp_index_num = 9
# for train_index, test_index in kf:
# # if not (index == max_sbp_index_num or index == max_dbp_index_num):
# # index += 1
# # continue
#
# if len(max_sbp_train_index) == 0:
# max_sbp_train_index = train_index
# if len(max_dbp_train_index) == 0:
# max_dbp_train_index = train_index
# if len(max_sbp_test_index) == 0:
# max_sbp_test_index = test_index
# if len(max_dbp_test_index) == 0:
# max_dbp_test_index = test_index
#
# rf = RegressionAlgorithm()
# # if index == max_sbp_index_num:
# rf.x_train = full_set_arr[train_index, :]
# rf.y_train = full_set_res[train_index]
# rf.x_test = full_set_arr[test_index, :]
# rf.y_test = full_set_res[test_index]
# # rf.train()
# rf.train_mbp()
# stk = StatsToolKits(rf.test(), full_set_res[test_index])
# cor = stk.get_pearson_corr()
# print('***************SBP CORR')
# print cor
# cor = cor[0]
# # rf.show_full_set_result('Tuning Model SBP Regression Result')
# sbp_corrs.append(cor)
# if cor > max_sbp_corr:
# max_sbp_index = index
# max_sbp_corr = cor
# max_sbp_train_index = train_index
# max_sbp_test_index = test_index
# rf.alter_type()
# rf.reset_model()
# rf.x_train = full_set_arr[train_index, :]
# rf.y_train = full_set_res[train_index, :]
# rf.x_test = full_set_arr[test_index, :]
# rf.y_test = full_set_res[test_index, :]
# rf.train()
# stk = StatsToolKits(rf.test(), full_set_res[test_index, 1])
# cor = stk.get_pearson_corr()
# cor = cor[0]
# # rf.show_full_set_result('Tuning Model DBP Regression Result')
# dbp_corrs.append(cor)
# print('***************DBP CORR')
# print(cor)
# if cor > max_dbp_corr:
# max_dbp_index = index
# max_dbp_corr = cor
# max_dbp_train_index = train_index
# max_dbp_test_index = test_index
# index += 1
index = 0
for train_index, test_index in kf:
if index == 1:
max_sbp_train_index = train_index
max_sbp_test_index = test_index
break
else:
index += 1
continue
print("sbp max index:" + str(max_sbp_index) + " ")
print(sbp_corrs)
print("dbp max index:" + str(max_dbp_index) + " ")
print(dbp_corrs)
rf = RegressionAlgorithm()
rf.x_train = full_set_arr[max_sbp_train_index, :]
rf.y_train = full_set_res[max_sbp_train_index]
rf.x_test = full_set_arr[max_sbp_test_index, :]
rf.y_test = full_set_res[max_sbp_test_index]
rf.train_mbp()
cor = rf.test()
# rf.show_full_set_result("Best SBP Regression")
rf.show_mbp_full_set_result("Best SBP Regression", list(rf.y_test))
stk = StatsToolKits(cor, full_set_res[max_sbp_test_index])
cor = stk.get_pearson_corr()
print('***************SBP CORR')
print cor
# rf.reset_model()
# rf.alter_type()
# rf.x_train = full_set_arr[max_dbp_train_index, :]
# rf.y_train = full_set_res[max_dbp_train_index, :]
# rf.x_test = full_set_arr[max_dbp_test_index, :]
# rf.y_test = full_set_res[max_dbp_test_index, :]
# rf.train()
# rf.test()
# rf.show_full_set_result("Best DBP Regression")
exit()
#
root_path = '/mnt/code/matlab/data/csv-long/'
only_csv_files = [f for f in listdir(root_path) if isfile(join(root_path, f)) &
f.startswith('a') & f.endswith('.csv')]
type_sbp_nums = list([0] * 4)
type_dbp_nums = list([0] * 4)
full_set_arr = []
full_set_res = []
rf = RegressionAlgorithm()
for csv_file_name in only_csv_files:
arr, res = rf.read_file(root_path + csv_file_name)
if np.size(arr, 0) >= RegressionAlgorithm.minFullSetSize:
full_set_arr = np.concatenate((full_set_arr, arr.tolist()), 0) if len(full_set_arr) > 0 else arr.tolist()
full_set_res = np.concatenate((full_set_res, res.tolist()), 0) if len(full_set_res) > 0 else res.tolist()
rf.split_sets(arr, res)
rf.train()
rf.test()
bhs_type = rf.get_result_bhs_type()
type_sbp_nums[bhs_type] += 1
rf.save_predict_result(csv_file_name, get_title(rf), '/mnt/code/matlab/data/csv/' + 'pic-long/' +
BHSTypes.get_type_name(bhs_type))
rf.alter_type()
rf.reset_model()
rf.train()
rf.test()
bhs_type = rf.get_result_bhs_type()
type_dbp_nums[bhs_type] += 1
rf.save_predict_result(csv_file_name, get_title(rf), '/mnt/code/matlab/data/csv/' + 'pic-long/' +
BHSTypes.get_type_name(bhs_type))
# rf.display_and_save_predict_result(csv_file_name)
# input()
# rf.display_and_save_predict_result()
# rf.train()
# rf.test()
exit()
rf.split_sets(full_set_arr, full_set_res)
rf.train()
rf.test()
# rf.show_full_set_result(get_title(rf))
disp_stats_paras(rf)
rf.reset_model()
rf.alter_type()
rf.train()
rf.test()
disp_stats_paras(rf)
# rf.show_full_set_result(get_title(rf))
# rf_obj.train()
# rf_obj.test()
# rf_obj.show_predict_result()
# print(type_sbp_nums)
# print(type_dbp_nums)