-
Notifications
You must be signed in to change notification settings - Fork 2
/
util_prepa.py
341 lines (295 loc) · 13.5 KB
/
util_prepa.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
# -*- coding: utf-8 -*-
"""
Created on Sun Aug 2 19:54:21 2020
Utility functions for financial analysis to prepare X and Y matrices
1. get_model_data((dbInput, dbList, ric, horiz, drop_rows, triggerUp1 = 0, triggerLoss = 0, triggerUp2 = 0)) => return dataX, dataY
2. get_model_Y(dataset, horizon, triggerUp1 = 0, triggerLoss = 0, triggerUp2 = 0) => return dataY
3. get_droprows(dataset, num_rows) => return dataset Remove rows in dataset improper due to tech indicators (usually 50 to 200 rows)
4. get_cleandupli_dataset(dataset) => return dataset
5. get_model_cleanXset(X_raw, trigger) => return X_clean Remove columns where number of missing values exceeds a trigger
6. get_train_test_resize(dataX, dataY, proportionTrain) => return (X_train, y_train), (X_test, y_test)
7. get_features_ident_xgb(dataX, proportionTrain) => return (X_train, y_train), (X_test, y_test)
@author: JDE65
"""
import pandas as pd
import numpy as np
import sqlite3 as sq
import matplotlib.pyplot as plt
from itertools import chain
from statsmodels.tsa.arima_model import ARIMA
from statsmodels.tsa.stattools import acf
###--- 1. Drop unnecessary rows into a dataset + summary ---###
def get_droprows(dataset, num_rows, horiz):
dataset = dataset.drop(dataset.index[0:num_rows])
dataset = dataset.drop(dataset.index[(len(dataset)-horiz):])
return dataset
###--- 2. Clean up a dataset for duplicates & fill in NaN ---###
def get_cleandupli_dataset(dataset):
dataset = dataset.drop_duplicates() #remove duplicated entries
#dataset = dataset.fillna(method='ffill') # fill NaN forward with next valid observation - opposite method = bfill
return dataset
###--- 3. Cleanup X : remove row with missing values above a trigger ---###
def get_cleanXset(X_raw, trigger): # ONLY if only basic data but no enriched data
X_raw = X_raw.iloc[:,2:]
nanc = X_raw.isna().sum()
for i in range(len(nanc)):
ananc = nanc.iloc[i]/len(X_raw)
if ananc > trigger:
X_raw = X_raw.drop(columns = i)
X_clean = X_raw.T.reset_index(drop = True).T
return X_clean
def get_cleanXset_full(X_raw, trigger):
X_raw = X_raw.iloc[:,4:]
nanc = X_raw.isna().sum()
for i in range(len(nanc)):
ananc = nanc.iloc[i]/len(X_raw)
if ananc > trigger:
X_raw = X_raw.drop(columns = i)
X_clean = X_raw.T.reset_index(drop = True).T
return X_clean
###--- 4. Cleanup X : remove row with missing values above a trigger ---###
def get3D_create_dataset(dataX, dataY, seq_length, step):
Xs, ys = [], []
for i in range(0, len(dataX) - seq_length + 1, step):
v = dataX.iloc[i:(i + seq_length)].values
Xs.append(v)
ys.append(dataY.iloc[i + seq_length - 1])
return np.array(Xs), np.array(ys)
###--- 6. Split X and Y dataset into X_train X_test y_train y_test ---###
## 6.1 Base case - Price
def get_train_test_price(dataX, dataY, nn_start, nn_size, proportionTrain):
Xs = np.array(dataX)
ys = np.array(dataY)
X = Xs[:,:] # dataX with 'instrument', 'date', 'close', 'open', ...
res = ys[:,2] * 1 # array with effective return over the horizon
futClose = ys[:,3] * 1 # array with future close of the stock
y = futClose * 1
y[np.isnan(y)] = 0 # convert nan into 0 for y
X[np.isnan(X)] = 0 # convert nan into 0 for X
if nn_start > 0:
X = X[nn_start:, :]
y = y[nn_start:,]
res = res[nn_start:]
if X.shape[0] > nn_size: # apply the split on the lowest of lstmSize or X length
train_size = int(nn_size * proportionTrain)
else:
train_size = int(X.shape[0] * proportionTrain)
X_train = X[:train_size, :]
X_test = X[train_size: nn_size, :]
y_train = y[:train_size]
y_test = y[train_size: nn_size]
res_train = res[:train_size]
res_test = res[train_size: nn_size]
return (X_train, y_train), (X_test, y_test), (res_train, res_test)
##--- 6.2. Create sliding window and enrich dataX and dataY ---###
def get3D_train_test_price(dataX, dataY, nn_start, nn_size, proportionTrain):
X = dataX[:, :, :] # dataX with 'instrument', 'date', 'close', 'open', ...
res = dataY[:,2] * 1 # separating the effective return between train & test
futClose = dataY[:,3] * 1
y = futClose * 1 # y = close price => regression on predicting future price
if nn_start > 0:
X = X[nn_start:, :, :]
y = y[nn_start:,]
res = res[nn_start:]
if X.shape[0] > nn_size: # apply the split on the lowest of lstmSize or X length
train_size = int(nn_size * proportionTrain)
else:
train_size = int(X.shape[0] * proportionTrain)
X_train = X[:train_size, :, :]
X_test = X[train_size: nn_size, :, :]
y_train = y[:train_size]
y_test = y[train_size: nn_size]
res_train = res[:train_size]
res_test = res[train_size: nn_size]
return (X_train, y_train), (X_test, y_test), (res_train, res_test)
##--- 6.3 Split X and Y dataset into X_train X_test y_train y_test ---###
def get_train_test_return(dataX, dataY, nn_start, nn_size, proportionTrain, modeRC):
Xs = np.array(dataX)
ys = np.array(dataY)
X = Xs[:,:] # dataX with 'instrument', 'date', 'close', 'open', ...
res = ys[:,2] * 1 # array with effective return over the horizon
futClose = ys[:,3] * 1 # array with future close of the stock
y = res * 1
if modeRC == 'class': # if classification => if return >=0 then buy, else let
y[y >= 0] = 1
y[y < 0.1] = 0 # either 0 or -1 to emphasize the cost of error with custom loss function
y[np.isnan(y)] = 0 # convert nan into 0 for y
X[np.isnan(X)] = 0 # convert nan into 0 for X
if nn_start > 0:
X = X[nn_start:, :]
y = y[nn_start:,]
res = res[nn_start:]
if X.shape[0] > nn_size: # apply the split on the lowest of lstmSize or X length
train_size = int(nn_size * proportionTrain)
else:
train_size = int(X.shape[0] * proportionTrain)
X_train = X[:train_size, :]
X_test = X[train_size: nn_size, :]
y_train = y[:train_size]
y_test = y[train_size: nn_size]
res_train = res[:train_size]
res_test = res[train_size: nn_size]
return (X_train, y_train), (X_test, y_test), (res_train, res_test)
##--- 6.3 Split X and Y dataset into X_train X_test y_train y_test ---###
def get3D_train_test_return(dataX, dataY, nn_start, nn_size, proportionTrain, modeRC):
X = dataX[:, :, :] # dataX with 'instrument', 'date', 'close', 'open', ...
res = dataY[:,2] * 1 # identifying the effective return achieve by the investment
y = res * 1 # y = close price => regression on predicting future price
if modeRC == 'class': # if classification => if return >=0 then buy, else let
y[y >= 0] = 1
y[y < 0.1] = -1
y1 = y.reshape(len(y),1) * 1
y2 = y.reshape(len(y),1) * 1
y1[y1<1]=0
y2[y2>0]=0
y2[y2<0]=1
y = np.append(y1, y2, axis = 1)
nb_classes = 2
else:
nb_classes = 1
if nn_start > 0:
X = X[nn_start:, :, :]
y = y[nn_start:,]
res = res[nn_start:]
if X.shape[0] > nn_size: # apply the split on the lowest of lstmSize or X length
train_size = int(nn_size * proportionTrain)
else:
train_size = int(X.shape[0] * proportionTrain)
X_train = X[:train_size, :, :]
X_test = X[train_size: nn_size, :, :]
y_train = y[:train_size]
y_test = y[train_size: nn_size]
res_train = res[:train_size]
res_test = res[train_size: nn_size]
return (X_train, y_train), (X_test, y_test), (res_train, res_test, nb_classes)
##--- 6.4 MODE Multi-Class of train-test split for return based on yield for 1 EUR invested
def get3D_train_test_returnMC(dataX, dataY, nn_start, nn_size, proportionTrain, modeRC, trig_up=0, trig_down=0):
X = dataX[:, :, :] # dataX with 'instrument', 'date', 'close', 'open', ...
res = dataY[:,2] * 1 # identifying the effective return achieve by the investment
y = res * 1 # y = close price => regression on predicting future price
if modeRC == 'class': # if classification => if return >=0 then buy, else let
nb_classes = 2
if trig_up > 0:
y[y >= trig_up] = 2
y[(y>= 0) & (y<2)] = 1
nb_classes += 1
else:
y[y >= 0] = 1
if trig_down < 0:
y[y <= trig_down] = -2
y[(y < 0) & (y>-2)] = -1
nb_classes += 1
else:
y[y < 0.1] = -1
y1 = y.reshape(len(y),1) * 1
y2 = y.reshape(len(y),1) * 1
y1[y1<1]=0
y2[y2>0]=0
y2[y2<-1]=2
y2[y2<0]=1
if trig_up > 0:
y3 = y.reshape(len(y),1) * 1
y3[y3 < 2] = 0
y3[y3 > 1] = 1
y1[y3 > 0] = 0
if trig_down < 0:
y4 = y.reshape(len(y),1) * 1
y4[y4 > -2] = 0
y4[y4 < -1] = 1
y2[y4 > 0] = 0
yt = np.append(y1, y2, axis = 1)
if trig_up > 0:
yt = np.append(yt, y3, axis = 1)
if trig_down < 0:
yt = np.append(yt, y4, axis = 1)
y = yt * 1
else:
nb_classes = 1
if nn_start > 0:
X = X[nn_start:, :, :]
y = y[nn_start:,]
res = res[nn_start:]
if X.shape[0] > nn_size: # apply the split on the lowest of lstmSize or X length
train_size = int(nn_size * proportionTrain)
else:
train_size = int(X.shape[0] * proportionTrain)
X_train = X[:train_size, :, :]
X_test = X[train_size: nn_size, :, :]
y_train = y[:train_size]
y_test = y[train_size: nn_size]
res_train = res[:train_size]
res_test = res[train_size: nn_size]
return (X_train, y_train), (X_test, y_test), (res_train, res_test, nb_classes)
###--- 7. Scale X_train X_test on X_train ---###
## 7.1
def get_scaleX(X_train, X_test):
train_mean = np.mean(X_train, axis = 0)
train_std = np.std(X_train, axis = 0)
X_train = (X_train - train_mean) / train_std
X_test = (X_test - train_mean) / train_std
return (X_train, X_test), (train_mean, train_std)
def get_minmaxscaleX(X_train, X_test):
train_mean = np.mean(X_train, axis = 0)
train_minmax = np.amax(X_train, axis = 0) - np.amin(X_train, axis = 0)
X_train = (X_train - train_mean) / train_minmax
X_test = (X_test - train_mean) / train_minmax
return (X_train, X_test), (train_mean, train_minmax)
## 7.2 Scaling for 3D dataset (LSTM, CNN, ResNet, ...)
def get3D_scaleX(X_train, X_test):
train_mean = np.mean(X_train, axis = 0)
train_mean = np.mean(train_mean, axis = 0)
train_std = np.std(X_train, axis = 0)
train_std = np.mean(train_std, axis = 0)
X_train = X_train - train_mean.T
X_train = X_train / train_std.T
X_test = X_test - train_mean.T
X_test = X_test / train_std.T
return (X_train, X_test), (train_mean, train_std)
def get3D_minmaxscaleX(X_train, X_test):
train_mean = np.mean(X_train, axis = 0)
train_mean = np.mean(train_mean, axis = 0)
train_minmax = np.amax(X_train, axis = 0) - np.amin(X_train, axis = 0)
train_minmax = np.mean(train_minmax, axis = 0)
X_train = X_train - train_mean.T
X_train = X_train / train_minmax.T
X_test = X_test - train_mean.T
X_test = X_test / train_minmax.T
return (X_train, X_test), (train_mean, train_minmax)
###--- 8. Split X matrix into X & Y with train & test for xgb_validation ---###
def get_features_ident_xgb(dataX, proportionTrain):
y = dataX['close'] # y is the closing price
X = dataX.iloc[:,3:] # X are all other elements of the X matrix
train_size = int(X.shape[0] * proportionTrain)
X_train = X.iloc[:train_size]
X_test = X.iloc[train_size:]
y_train = y.iloc[:train_size]
y_test = y.iloc[train_size:]
return (X_train, y_train), (X_test, y_test)
###--- 11. To be developped ---###
def plot_technical_indicators(dataset, last_days):
plt.figure(figsize=(16, 10), dpi=100)
shape_0 = dataset.shape[0]
xmacd_ = shape_0-last_days
dataset = dataset.iloc[-last_days:, :]
x_ = range(3, dataset.shape[0])
x_ =list(dataset.index)
# Plot first subplot
plt.subplot(2, 1, 1)
plt.plot(dataset['mav05'],label='MA 5', color='g',linestyle='--')
plt.plot(dataset['price'],label='Closing Price', color='b')
plt.plot(dataset['mav20'],label='MA 20', color='r',linestyle='--')
plt.plot(dataset['bollUpperband'],label='Upper Band', color='c')
plt.plot(dataset['bollLowerband'],label='Lower Band', color='c')
plt.fill_between(x_, dataset['bollLowerband'], dataset['upper_band'], alpha=0.35)
plt.title('Technical indicators for the dataset - last {} days.'.format(last_days))
plt.ylabel('EUR')
plt.legend()
# Plot second subplot
plt.subplot(2, 1, 2)
plt.title('MACD')
plt.plot(dataset['macd'],label='MACD', linestyle='-.')
plt.hlines(15, xmacd_, shape_0, colors='g', linestyles='--')
plt.hlines(-15, xmacd_, shape_0, colors='g', linestyles='--')
plt.plot(dataset['log_momentum'],label='Momentum', color='b',linestyle='-')
plt.legend()
plt.show()