forked from Tingji2419/MSS
-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathgenerate_series.py
164 lines (122 loc) · 5.59 KB
/
generate_series.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
from utils import set_seed
import numpy as np
import random
from numpy.lib.stride_tricks import sliding_window_view
def constant_signal(am=None, length=120, freq=0.02, noise=True):
am = random.randint(-10, 10) / 10 if not am else am
if noise:
return am * np.ones(length) + np.random.normal(0, 1, length) * 0.05
return am * np.ones(length)
def linear_signal(am=None, length=120, freq=0.02, noise=True):
am = random.randint(-5, 5) / 10 if not am else am
am = am / length
cm = random.randint(-5, 5) / 10
x = np.arange(0, length * freq, freq)
y = am * (x) + cm
# if noise:
# return y + np.random.normal(0, 1, length) * 0.005
return y
def stair_signal(am=None, length=120, freq=0.02, noise=True):
am = random.randint(-5, 5) / 100 if not am else am
bm = random.randint(30, length) // 10
x = np.arange(0, bm) * am
x = np.repeat(x, length // bm + 1)
return x[:length] + np.random.normal(0, 1, length) * 0.05
def sawtooth_signal(am=None, length=120, freq=0.02, noise=True):
am = random.randint(-5, 5) / 10 if not am else am
bm = random.randint(30, length) // 10
x = np.arange(0, bm) * am
x = np.concatenate([x] * (length // bm + 1))
return x[:length] + np.random.normal(0, 1, length) * 0.05
def square_signal(am=None, length=120, freq=0.02, noise=True):
am = random.randint(-5, 5) / 10 if not am else am
bm = random.randint(30, length) // 4
x = np.ones(length) * am
for index in range(length):
if (index // bm) % 2 == 0:
x[index] = -am
return x + np.random.normal(0, 1, length) * 0.05
def sin_signal(am=None, length=120, freq=0.02, noise=True):
am = random.randint(-10, 10) / 10 if not am else am
if noise:
noise = np.random.normal(0, 1, length)
return am * np.sin(np.arange(0, length * freq, freq) * 5) + noise * 0.05
return am * np.sin(np.arange(0, length * freq, freq) * 5)
def sin_signal_mix_sawtooth(am=None, length=120, freq=0.02, noise=True):
return sin_signal(am, length, freq, noise) * sawtooth_signal(0.01, length, freq, noise)
def sin_signal_mix_square(am=None, length=120, freq=0.02, noise=True):
return sin_signal(am, length, freq, noise) * square_signal(am, length, freq, noise)
def sin_signal_mix_cos(am=None, length=120, freq=0.02, noise=True):
return sin_signal(am, length, freq, noise) * cos_signal(0.01, length, freq, noise)
def cos_signal(am=None, length=120, freq=0.02, noise=True):
am = random.randint(-10, 10) / 10 if not am else am
if noise:
noise = np.random.normal(0, 1, length)
return am * np.sin(np.arange(0, length * freq, freq) * 5) + noise * 0.05
return am * np.sin(np.arange(0, length * freq, freq) * 5)
# def cos_signal_mix(am=None, length=120, freq=0.02, noise=True):
# return cos_signal(am, length, freq, noise) * sawtooth_signal(am, length, freq, noise)
SIGNAL = {
'constant': constant_signal,
# 'constant2': constant_signal,
'linear': linear_signal,
'sawtooth': sawtooth_signal,
'stair': stair_signal,
'sin': sin_signal,
'cos': cos_signal,
'square': square_signal,
'sin2': sin_signal_mix_sawtooth,
# 'cos2': cos_signal_mix,
'sin3': sin_signal_mix_square,
'sin4': sin_signal_mix_cos,
}
set_seed(42)
class Artifical_Signal():
'''
Generate artifical signals to train discriminant models
'''
def __init__(self, num=5000, windows_size=250) -> None:
self.num = num
self.ws = windows_size
self.cps = None
self.series = np.zeros(self.num + self.ws)
self.generate_change_points()
def generate_change_points(self):
'''
Function: Generate num/2 change points to make balance
Frequency: 50Hz
'''
ratio_y = 0.5
min_action = int(self.ws) * 1
if self.cps == None:
self.cps = sorted(random.sample(list(np.arange(0, self.num, min_action)), int(ratio_y *self.num / min_action)))
return self.cps
def generate_signal(self):
# 换成连续一条，然后滑动窗口
signal_choices = [i for i in SIGNAL.values()]
for pre, cur in zip(self.cps[:-1], self.cps[1:]):
cur_signal = random.choice(signal_choices)
self.series[pre: cur] += cur_signal(am=None, length=cur-pre, freq=0.02, noise=True)
labels = np.zeros(self.num + self.ws)
labels[self.cps] = 1
self.labels = [int(sum(i) > 0) for i in sliding_window_view(labels, self.ws)]
return sliding_window_view(self.series, self.ws)[1:], self.labels[1:]
# self.generate_change_points()
# choices = [i for i in SIGNAL.values()]
# for index, i in enumerate(self.cps):
# self.series[index * self.ws: index * self.ws + i] = random.choice(choices)(am=None, length=self.ws, freq=0.02, noise=True)
# self.series[index * self.ws + i + 1:] = random.choice(choices)(am=None, length=self.es, freq=0.02, noise=True)
# for index, _ in range(self.num / 2):
# self.series[index * self.ws: (index + 1) * self.ws] = random.choice(choices)(am=None, length=self.ws, freq=0.02, noise=True)
# return self.series, self.cps
if __name__ == '__main__':
import matplotlib.pyplot as plt
# a = linear_signal(am=None, length=120, freq=0.02, noise=True)
# plt.plot(a)
# plt.savefig('2.png')
dataset = Artifical_Signal(num=50000, windows_size=200)
X, y = dataset.generate_signal()
print('Totoal:', len(y))
print('1 in label:', sum(y) / len(y))
plt.plot(dataset.series[:30000])
plt.savefig('1.png')