-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathpackage.py
212 lines (148 loc) · 8.13 KB
/
package.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
import pandas as pd
import numpy as np
from sklearn.linear_model import LinearRegression, Ridge, Lasso
from sklearn import metrics
from sklearn import ensemble
from sklearn.model_selection import GridSearchCV, RandomizedSearchCV
def return_tot(df, date, value_beg, value_end):
period_beg = df.date.min()
period_end = df.date.max()
price_beg = df.loc[df.date == period_beg, value_beg].values[0]
price_end = df.loc[df.date == period_end, value_end].values[0]
ret = (price_end-price_beg)/ price_beg
return ret
class Model:
def __init__(self,df):
self.df = df
print('object instanciated')
def train(self, df=None, col_period=None , train_window=4 , test_window=1, test_gap = 0, expanding=False):
test_train = []
# test_train = {}
periods = df[col_period].unique().tolist()
periods.sort()
if expanding == False:
for i,j in enumerate(periods):
if i < len(periods)-train_window-1:
train_beg = j
train_beg_idx = periods.index(train_beg)
train_end_idx = train_beg_idx+ train_window
train_end = periods[train_end_idx]
test_beg_idx = train_end_idx+test_gap
test_end_idx = test_beg_idx+test_window
# train = periods[train_idx]
train_periods = periods[train_beg_idx: train_end_idx]
test_periods = periods[test_beg_idx: test_end_idx]
df_train = df.loc[df[col_period].isin(train_periods)]
df_test = df.loc[df[col_period].isin(test_periods)]
test_train.append((df_train, df_test))
# test_train[i] = {"train": df_train , "test": df_test}
else:
for i,j in enumerate(periods):
if i < len(periods)-train_window-1:
train_beg = periods[0]
train_beg_idx = periods.index(train_beg)
train_end_idx = train_beg_idx+ train_window+i
train_end = periods[train_end_idx]
test_beg_idx = train_end_idx+test_gap
test_end_idx = test_beg_idx+test_window
# train = periods[train_idx]
train_periods = periods[train_beg_idx: train_end_idx]
test_periods = periods[test_beg_idx: test_end_idx]
df_train = df.loc[df[col_period].isin(train_periods)]
df_test = df.loc[df[col_period].isin(test_periods)]
test_train.append((df_train, df_test))
# test_train[i] = {"train": df_train , "test": df_test}
return test_train
def skpredict(self, df_train, df_test, skmodel, cols_x, cols_y, printstat=True):
regr = skmodel
X_train = df_train[cols_x]
Y_train = df_train[cols_y]
X_test = df_test[cols_x]
Y_test = df_test[cols_y]
regr.fit(X_train, Y_train)
predict_train = regr.predict(X_train)
predict_test = regr.predict(X_test)
df_train['predict'] = predict_train
df_train['MSE'] = (np.array(df_train[cols_y]) - predict_train)**2
df_test['predict'] = predict_test
df_test['MSE'] = (np.array(df_test[cols_y]) - predict_test)**2
mse_train = metrics.mean_squared_error(Y_train, predict_train)
mae_train = metrics.mean_absolute_error(Y_train, predict_train)
r2_train = metrics.r2_score(Y_train, predict_train)
mse_test = metrics.mean_squared_error(Y_test, predict_test)
mae_test = metrics.mean_absolute_error(Y_test, predict_test)
stat_train = {"mse": mse_train, "mae": mae_train, "r2": r2_train}
stat_test = {"mse": mse_test, "mae": mae_test}
if printstat==True:
print(f'train stat: {stat_train}')
print(f'test stat: {stat_test}')
# print(f'MSE Score manually calulated: {np.mean((np.array(df_final_vf[cols_y]) - predict)**2)}')
return df_train, df_test, stat_train, stat_test
def skpredict_window(self, df, skmodel, cols_x, cols_y, col_period , train_window=4 , test_window=1, test_gap = 0, expanding=False, print_iter=False):
regr = skmodel
train_test = self.train(df, col_period, train_window , test_window, test_gap, expanding)
data_train = []
data_test = []
stat_train_times = []
stat_test_times = []
for i , j in enumerate(train_test):
df_train = j[0]
df_test = j[1]
df_train, df_test, stat_train, stat_test = self.skpredict(df_train, df_test, regr, cols_x, cols_y, printstat=False)
stat_train.update({"window": i, "date": df_train[col_period].unique().tolist()[0] + "-"+ df_train[col_period].unique().tolist()[-1]})
stat_test.update({"window": i, "date": df_test[col_period].unique().tolist()[0] + "-"+ df_test[col_period].unique().tolist()[-1]})
df_train['iteration'] = i
df_test['iteration'] = i
data_train.append(df_train)
data_test.append(df_test)
stat_train_times.append(stat_train)
stat_test_times.append(stat_test)
if print_iter != False:
print(f'train stat: {stat_train}')
print(f'test stat: {stat_test}\n')
df_train_conso = pd.concat(data_train)
df_test_conso = pd.concat(data_test)
mse_train_all = df_train_conso.MSE.mean()
mse_test_all = df_test_conso.MSE.mean()
dict_stat = {"MSEtrain": mse_train_all, "MSEtest": mse_test_all}
print(f'Average MSE train: {mse_train_all}')
print(f'Average MSE test: {mse_test_all}')
return df_train_conso , df_test_conso, stat_train_times, stat_test_times , dict_stat
def backtest(self, df, value, col_return, new_col_name):
"""
"""
## produce array for t and t-1 indexing
period_arr = np.empty([0,2])
period_idx = [i for i in df.index]
for i,j in enumerate(period_idx):
if i ==0:
x, y = j, np.nan
else:
x,y = j, period_idx[i-1]
arr = np.array([[x,y]])
period_arr = np.append(period_arr, arr, axis=0)
## back test results of strategy
price_beg = df.loc[period_arr[0,0]][value]
for i in df.index:
period_prev = period_arr[period_arr[:,0] ==i][:,1][0]
return_predict = df.loc[i][col_return]
if i == df.index.min():
pass
else:
price_prev = df.loc[period_prev][value]
return_strat = return_predict * price_beg
price_end = price_beg + return_strat
# print(i, price_prev, price_now, return_act, return_predict, predict, return_strat ,price_beg, price_end)
df.loc[i, new_col_name] = price_end
price_beg = price_end
beg_val = df[df.index==df.index.min()][value][0]
df.loc[df.index==df.index.min(), new_col_name] = beg_val
return df
def skbacktest(self, df, skmodel, cols_x, cols_y, col_period , col_value, train_window=4 , test_window=1, test_gap = 0, expanding=False, print_iter=False):
df_train_conso , df_model, stat_train_times, stat_test_times , dict_stat = self.skpredict_window(df, skmodel, cols_x, cols_y, col_period , train_window, test_window, test_gap, expanding, print_iter)
df_model.loc[df_model['predict'] > 0, 'strat_return'] = df_model[cols_y]
df_model.loc[df_model['predict'] < 0, 'strat_return'] = -df_model[cols_y]
df_model = self.backtest(df_model, 'value' , 'strat_return', 'value_strat')
# beg_val = df_model[df_model['date']==df_model['date'].min()][col_value][0]
# df_model.loc[df_model['date']==df_model['date'].min(), 'value_strat'] = beg_val
return df_model, dict_stat