/
eval_reg.py
207 lines (154 loc) · 6.75 KB
/
eval_reg.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
from Meta_Regression import *
import numpy as np
import pandas as pd
from mpl_toolkits.mplot3d import Axes3D
from sklearn.linear_model import LinearRegression
import matplotlib.pyplot as plt
from Linear_regression import linear_regression
from utils import data_preprocessing, evaluation
import matplotlib.ticker as mtick
from torchsummary import summary
import warnings
warnings.filterwarnings("ignore")
def regress(trainX, trainY, test, dep_var, performance):
methods = [LinearRegression]
print(np.asarray(trainY))
for method in methods:
sk_linear_reg = method().fit(np.asarray(trainX), np.asarray(trainY))
pred_reg_t = sk_linear_reg.predict(np.asarray(test.drop([dep_var], axis=1)))
rmse, mae, smape = evaluation(test[dep_var], pred_reg_t)
performance['Data'].append(data_name)
performance['Method'].append('OLS')
performance['Index'].append(k)
performance['RMSE'].append(rmse)
performance['MAE'].append(mae)
performance['sMAPE'].append(smape)
return performance
dep_vars= ['CO2']
data_names =['CO_2']
# Hyper-parameters
hidden_size = [16]#, 32, 64, 32, 10]
num_output = 1
num_epochs = 5000
learning_rate = 0.005
patience = 1000
batch_size = 16
k=0
for data_name, dep_var in zip(data_names, dep_vars):
data = pd.read_csv('data/%s.csv' %data_name)
data=data.dropna()
data = data.set_index('NAME', drop=True)
data_name='CO_3'
# add intercept
data['GNP'] = np.log(data['GNP'])
#data['GNP^2'] = (np.square(data['GNP']))
data['intercept'] = 1
data[dep_var]=np.log(data[dep_var])
performance = {'Data': [], 'Method': [], 'Index': [],
'RMSE': [], 'MAE': [], 'sMAPE': []}
train = data.loc[data['origin']==0, ]
test = data.loc[data['origin']==1, ]
train = train.drop(['origin'], axis=1)
test = test.drop(['origin'], axis=1)
trainY = train[dep_var]
#testY = test[dep_var]
# print(np.mean(trainY), np.std(trainY))
# print(np.mean(testY), np.std(testY))
#
# plt.style.use('seaborn-white')
# kwargs = dict(histtype='stepfilled', alpha=0.5, density=True, bins=8, ec="k")
#
# plt.hist(trainY, **kwargs, label='2006: mean=0.73, sigma=1.53')
# plt.hist(testY, **kwargs, label='2016: mean=0.49, sigma=1.76')
# plt.xlabel('CO2 emission')
# plt.ylabel('Probability density')
# plt.title('Histogram of CO2 emission')
# plt.legend()
# plt.show()
trainX = train.drop([dep_var], axis=1)
#estimate linear regression
linear_reg = linear_regression(train, dep_var)
linear_model, sel_variables = linear_reg.regression()
print(linear_model.summary())
#exit('b')
input_size = len(sel_variables)
conf_int = linear_model.conf_int()
coef = linear_model.params
std = linear_model.bse
conf_int[0] = coef
conf_int[1] = std
train_set, val_set, test_set = data_preprocessing(train=train, test=test,
variables=sel_variables,
conf_int=conf_int, dep_var=dep_var)
meta_reg = MetaRegression(input_size=input_size,
hidden_size=hidden_size,
output_size=num_output)
#print(summary(meta_reg, [(1, 12), (1, 12), (1, 12), (1, 12)]))
if os.path.isfile('models/checkpoint_%s_%s.pt' % (data_name, k)):
meta_reg = training(meta_reg, train_set, val_set, epochs=num_epochs,
batch_size=batch_size, lr=learning_rate, data_name=data_name,
idx=k, patience=patience)
else:
meta_reg, train_loss, valid_loss = training(meta_reg, train_set, val_set, epochs=num_epochs,
batch_size=batch_size, lr=learning_rate, data_name=data_name,
idx=k, patience=patience)
fig = plt.figure(figsize=(10, 8))
plt.plot(range(1, len(train_loss) + 1), train_loss, label='Training Loss')
plt.plot(range(1, len(valid_loss) + 1), valid_loss, label='Validation Loss')
# find position of lowest validation loss
minposs = valid_loss.index(min(valid_loss)) + 1
plt.axvline(minposs, linestyle='--', color='r', label='Early Stopping Checkpoint')
plt.xlabel('epochs')
plt.ylabel('loss')
plt.grid(True)
plt.legend()
plt.tight_layout()
plt.savefig('reg/early_stop_%s_%s.png' %(data_name, k))
plt.close()
output, coeff, testY, p_value = predict(meta_reg, test_set)
output = output.detach().numpy()
coeff = coeff.detach().numpy()
testY = testY.detach().numpy()
p_value = p_value.detach().numpy()
# for i in range(2):
# plt.scatter(np.abs(np.asarray(testY)-np.asarray(output)),
# p_value[:,i], c='r', alpha=0.5)
# plt.show()
coef_pd = pd.DataFrame(data=coeff, columns=sel_variables)
conc_out_pd = pd.DataFrame(data=p_value, columns=sel_variables)
rmse, mae, smape = evaluation(testY, output)
performance['Data'].append(data_name)
performance['Method'].append('Meta_reg')
performance['Index'].append(k)
performance['RMSE'].append(rmse)
performance['MAE'].append(mae)
performance['sMAPE'].append(smape)
performance = regress(trainX=trainX, trainY=trainY, test=test,
dep_var=dep_var, performance=performance)
font = {'family': 'serif',
'color': 'black',
'weight': 'normal',
'size': 12,
}
#test['GNP^2']=test['GNP']**2
for var in sel_variables:
other = test[sel_variables].drop([var], axis=1)
for other_var in list(other):
fig = plt.figure(figsize=(8, 8))
ax = fig.add_subplot(111, projection='3d')
ax.scatter(test[var], other[other_var], coef_pd[var], c='darkred',
marker='o')
ax.set_xlabel(var, fontdict=font)
ax.set_ylabel(other_var, fontdict=font)
ax.set_zlabel('Effect of %s' %var, fontdict=font)
ax.zaxis.set_major_formatter(mtick.FormatStrFormatter('%.1f'))
#plt.title('The relationship between CO2 and GNP')
#plt.show()
plt.savefig('reg/scatter_%s_%s_%s_%s.png' % (data_name, var, other_var, k))
plt.close()
res = pd.DataFrame.from_dict(performance)
res.to_csv('reg/performance.csv', index=False, mode='a')
coef_pd.index = test.index
conc_out_pd.index = test.index
res_df = pd.concat((test, coef_pd, conc_out_pd), axis=1)
res_df.to_csv('reg/df_%s.csv' %data_name, index=True)