-
Notifications
You must be signed in to change notification settings - Fork 1
/
Copy pathData_processing.py
executable file
·168 lines (142 loc) · 6.04 KB
/
Data_processing.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
import torch
import numpy as np
from sklearn.preprocessing import MinMaxScaler, StandardScaler
import warnings
import joblib
from generate_data import generate_data
warnings.filterwarnings("ignore")
class pre_process():
'''
Prepare the spectra (data) for training the neural network:
1- standardization of the data.
2- Reshaping the data as input for the neural network
3- Splitting the data for training, validation, and testing (0.7,0.2,0.1)
4- Splitting test data for curve fitting
'''
def __init__(self, scale=True, scale_type='stand', device=0,
New_Training=False,
data=None,
Neual_Network='net1'):
'''
args:
scale (Bool): Data is scaled if the value is True
scale_type (str): scaling type; normalization or min-max scaling
device (int): cuda decive number if available
data (numpy): the traing data
Neural_Network 'str': the choice of the neural network, net1 or net2 (see Module NN.py)
'''
self.w = data[0]
self.Z_real = data[1]
self.Z_imag = data[2]
self.Z_real_wide = data[3]
self.Z_imag_wide = data[4]
self.onehot_vector = data[5]
self.Binary_Matrix = data[6]
self.degs = data[7]
self.scale = scale
self.scale_type = scale_type
self.model = Neual_Network
self.New_Training = New_Training
self.generator = generate_data()
self.ind_ub = self.generator.ind_ub
def process(self):
'''
the pre-processing function.
returns:
1- Tuple containing the training the validation set
2- Tuple of test data
3- Tuple of data for the curve fitting part
'''
Norm_type = self.scale_type
hot_vector = self.onehot_vector
Binary_Matrix = self.Binary_Matrix
Z_real = self.Z_real
Z_imag = self.Z_imag
degs = self.degs
Z_real_og = np.copy(Z_real)
Z_imag_og = np.copy(Z_imag)
ind_ub = self.ind_ub
Z_imag = (Z_imag + ind_ub*2*np.pi*self.w.max())
Z_real, Z_imag = np.log(Z_real), np.log(Z_imag)
if self.New_Training:
if Norm_type == 'minmax':
scaler_real = MinMaxScaler()
scaler_imag = MinMaxScaler()
elif Norm_type == 'stand':
scaler_real = StandardScaler()
scaler_imag = StandardScaler()
scaler_real.fit(Z_real)
scaler_imag.fit(Z_imag)
Z_real = scaler_real.transform(Z_real)
Z_imag = scaler_imag.transform(Z_imag)
joblib.dump(scaler_real, 'scaler_real.gz')
joblib.dump(scaler_imag, 'scaler_imag.gz')
else:
scaler_real = joblib.load('scaler_real.gz')
scaler_imag = joblib.load('scaler_imag.gz')
Z_real = scaler_real.transform(Z_real)
Z_imag = scaler_imag.transform(Z_imag)
input_ = np.array((Z_real, Z_imag, Z_real_og, Z_imag_og))
input_ = np.swapaxes(input_, 0, 1)
if self.model == 'net1':
input_ = input_.reshape(input_.shape[0], 1,
input_.shape[1], input_.shape[2])
split = 0.8
train_data = input_[:int(input_.shape[0]*split)].astype(float)
test_data = input_[int(input_.shape[0]*split):].astype(float)
hot_vec_train = hot_vector[:int(input_.shape[0]*split)]
hot_vec_test = hot_vector[int(input_.shape[0]*split):]
binary_train = Binary_Matrix[:int(input_.shape[0]*split)]
binary_test = Binary_Matrix[int(input_.shape[0]*split):]
Z_real_test = Z_real_og[int(input_.shape[0]*split):]
Z_imag_test = Z_imag_og[int(input_.shape[0]*split):]
Z_real_test_wide = self.Z_real_wide[int(input_.shape[0]*split):]
Z_imag_test_wide = self.Z_imag_wide[int(input_.shape[0]*split):]
degs_test = degs[int(input_.shape[0]*split):].astype(float)
train_set = torch.utils.data.TensorDataset(
torch.from_numpy(train_data),
torch.from_numpy(hot_vec_train),
torch.from_numpy(binary_train))
test_set = torch.utils.data.TensorDataset(
torch.from_numpy(test_data),
torch.from_numpy(hot_vec_test),
torch.from_numpy(binary_test))
split_train = 0.7
split_val = 0.2
split_test = 0.1
num_samples = input_.shape[0]
train_end = int(num_samples * split_train)
val_end = train_end + int(num_samples * split_val)
train_data = input_[:train_end].astype(float)
val_data = input_[train_end:val_end].astype(float)
test_data = input_[val_end:].astype(float)
hot_vec_train = hot_vector[:train_end]
hot_vec_val = hot_vector[train_end:val_end]
hot_vec_test = hot_vector[val_end:]
binary_train = Binary_Matrix[:train_end]
binary_val = Binary_Matrix[train_end:val_end]
binary_test = Binary_Matrix[val_end:]
Z_real_test = Z_real_og[val_end:]
Z_imag_test = Z_imag_og[val_end:]
Z_real_test_wide = self.Z_real_wide[val_end:]
Z_imag_test_wide = self.Z_imag_wide[val_end:]
degs_test = degs[val_end:].astype(float)
# Create datasets
train_set = torch.utils.data.TensorDataset(
torch.from_numpy(train_data),
torch.from_numpy(hot_vec_train),
torch.from_numpy(binary_train)
)
val_set = torch.utils.data.TensorDataset(
torch.from_numpy(val_data),
torch.from_numpy(hot_vec_val),
torch.from_numpy(binary_val)
)
test_set = torch.utils.data.TensorDataset(
torch.from_numpy(test_data),
torch.from_numpy(hot_vec_test),
torch.from_numpy(binary_test)
)
return (train_set, val_set), \
(test_data, Z_real_test, Z_imag_test, degs_test, binary_test), \
(Z_real_test_wide, Z_imag_test_wide)