-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathperformance.py
161 lines (127 loc) · 5.18 KB
/
performance.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
# -*- coding: utf-8 -*-
"""
Created on Mon Jul 12 09:55:55 2021
@author: J Xin
"""
import numpy as np
import pandas as pd
import empyrical as ep
import matplotlib.pyplot as plt
import requests
from datetime import timedelta
from sklearn.metrics import roc_auc_score, plot_roc_curve,roc_curve, auc, accuracy_score
from my_pypackage.datadownloader import DataDownloader
from my_pypackage.preprocessors import FeatureEngineer
class Model_Performance(object):
def __init__(self, sec_code, signal_df, model_name, period):
"""
Parameters
----------
sec_code : TYPE
DESCRIPTION.
signal_df : DataFrame
columns list = ['date', 'signal', 'probability']
model_name : TYPE
DESCRIPTION.
Returns
-------
None.
"""
self.sec_code = sec_code
self.signal_df = signal_df
self.model_name = model_name
self.period = period
def _get_open_price(self):
signal_df = self.signal_df.reset_index()
signal_df = signal_df.sort_values(by='date',ascending=True)
start_date = pd.Timestamp(signal_df['date'].values[0])
end_date = pd.Timestamp(signal_df['date'].values[-1]) + timedelta(days=7)
downloader = DataDownloader(str(start_date), str(end_date), self.sec_code)
res = downloader.fetch_open_price()
return res
def _get_signal_merged_with_price(self):
# signal = self.signal_df[['signal']]
signal = self.signal_df.copy()
# signal = signal.set_index('date')
signal = FeatureEngineer.resample_to_business_day(signal)
price = self._get_open_price()
merged = pd.merge(signal, price, on='date', how='outer').fillna(method='ffill')
merged = merged.sort_values(by='date', ascending=True)
return merged
@classmethod
def get_position(self, data):
"""
Parameters
----------
data : DataFrame
index = 'date'
column list = ['signal', 'probability', 'open'].
Returns
-------
None.
"""
df = data.copy()
df['position'] = df['signal'].shift(1)
return df
@classmethod
def get_daily_return(self, data):
"""
Parameters
----------
data : DataFrame
index = 'date'
colum list = ['siganl', 'probability', 'open', 'position'].
Returns
-------
df : TYPE
DESCRIPTION.
"""
df = data.copy()
df.loc[:, 'price_change_pct'] = df.loc[:, 'open'].pct_change()
df.loc[:, 'turnover'] = df['position'].diff().fillna(0)
df.loc[:, 'daily_return_before_cost'] = df.price_change_pct * df['position'].shift(1).fillna(0)
df.loc[:, 'daily_return'] = (df.loc[:, 'daily_return_before_cost'] - df.loc[:, 'turnover'].abs() * 0.0003).fillna(0)
# df.pop('turnover')
return df
def BackTestStats(self, data):
df = data.copy()
df_copy = df.copy()
df_copy = df_copy.loc[df_copy.loc[:, 'daily_return'] != 0, :]
sharpe = ep.sharpe_ratio(df['daily_return'].values, period=self.period)
annual_return = ep.annual_return(df['daily_return'].values, period=self.period)
annual_volatility = ep.annual_volatility(df['daily_return'].values, period=self.period)
win_ratio = (np.sign(df['daily_return']) == 1).sum() / len(df['daily_return'].values)
auc_score = roc_auc_score(np.sign(df_copy['daily_return'].values)[1:], df_copy['probability'].fillna(0).shift(1).values[1:], average=None)
max_drawdown = ep.max_drawdown(df['daily_return'].values)
stats = {'Annual return': [annual_return], 'Annual volatility': [annual_volatility],
'Sharpe ratio': [sharpe], 'Max drawdown': [max_drawdown],
'Win Rate': [win_ratio], 'AUC': [auc_score]}
res = pd.DataFrame(stats)
res.index = [self.model_name]
return res
def BackTestPlot(self, df):
df.loc[:, 'net_value'] = (1 + df.loc[:, 'daily_return'].values).cumprod()
fig = plt.figure()
ax1 = fig.add_subplot(111)
x = df.index
y = df.loc[:, 'net_value']
w = df.loc[:, 'signal']
ax1.plot(x, y, color='r', label='net_value')
ax2 = ax1.twinx()
ax2.bar(x, w, color='orange', label='position')
plt.xlabel('date')
ax1.legend(loc='upper left')
ax2.legend(loc='upper right')
plt.title('_'.join((self.model_name, 'positions')))
plt.grid()
plt.savefig('.'.join((self.model_name + '_positions', 'png')))
plt.close()
# plt.show()
return
def run(self):
df = self._get_signal_merged_with_price()
position = self.get_position(df)
ret = self.get_daily_return(position)
stat = self.BackTestStats(ret)
self.BackTestPlot(ret)
return ret, stat