-
Notifications
You must be signed in to change notification settings - Fork 0
/
Part 3 - Regression.py
297 lines (207 loc) · 8.27 KB
/
Part 3 - Regression.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
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
# -*- coding: utf-8 -*-
"""
Created on Wed Mar 18 10:25:21 2020
@author: apost
"""
# -*- coding: utf-8 -*-
"""
Created on Fri Feb 28 13:45:06 2020
@author: apost
"""
import pandas as pd
import numpy as np
from sklearn.model_selection import train_test_split
import matplotlib.pyplot as plt
from sklearn.metrics import r2_score
from sklearn import linear_model
from sklearn.metrics import mean_squared_error
from ml_metrics import rmse
from ml_metrics import mae
from sklearn.preprocessing import PolynomialFeatures
from sklearn.linear_model import Ridge
from sklearn.linear_model import Lasso
def scores(actuals, predicteds):
rmses = rmse(actual = actuals, predicted = predicteds)
mses = mean_squared_error(actuals,predicteds)
maes = mae(actual = actuals, predicted = predicteds)
r2s = r2_score(actuals,predicteds)
return rmses,maes,r2s,mses
#import the dataset
data = pd.read_csv('FuelConsumptionCo2.csv')
#keep the lines that play a role in co2emissions
cols = ['ENGINESIZE','CYLINDERS', 'FUELTYPE', 'FUELCONSUMPTION_CITY','FUELCONSUMPTION_HWY','FUELCONSUMPTION_COMB','CO2EMISSIONS']
df = data[cols]
print(df['FUELTYPE'].unique())
#factorize, like dummies but in a single column, converts letter categories to number categoriess
df['FUELTYPE'] = pd.factorize(df['FUELTYPE'])[0]
df['FUELTYPE'] +=1 #because we need no zeros for the log transformation
print(df['FUELTYPE'].unique())
#log transformation to increase accuracy
df = np.log(df)
print()
print()
#------------------------------ Simple Linear Regression -------------------
print('------------------Simple Linear Regression: ')
print()
#linear model for how engine size affects co2 emmissions
model = linear_model.LinearRegression()
xtrain,xtest,ytrain,ytest = train_test_split(df[['ENGINESIZE']],df[['CO2EMISSIONS']],test_size = 0.3)
model.fit(xtrain,ytrain)
SR_trainpred = np.array(model.predict(xtrain))
SR_testpred = np.array(model.predict(xtest))
print('Regression Line for engine size / CO2 emissions:')
#regression line for the predicted in test
plt.scatter(xtest, ytest, color='maroon')
plt.plot(xtest, SR_testpred, color='cyan', linewidth=2)
plt.show()
print()
#print metrics for the simple linear regression model
SR_metrics_train = scores(ytrain,SR_trainpred)
SR_metrics_test = scores(ytest, SR_testpred)
print('Train mse: ',SR_metrics_train[3])
print('Test mse: ',SR_metrics_test[3])
print('Train rmse: ',SR_metrics_train[0])
print('Test rmse: ',SR_metrics_test[0])
print('Train mae: ',SR_metrics_train[1])
print('Test mae: ',SR_metrics_test[1])
print('Train R2: ',SR_metrics_train[2])
print('Test R2: ',SR_metrics_test[2])
print()
print()
#---------------------------------Polynomial Regression-----------------------
print('------------Polynomial Regression: ')
print()
#split the dataset for fitting
X_train, X_test, Y_train, Y_test = train_test_split(df[['ENGINESIZE']],df[['CO2EMISSIONS']], test_size=0.3)
#regression model
poly = PolynomialFeatures(degree=2)
#regr = linear_model.LinearRegression()
polyTR=poly.fit_transform(X_train)
polyTE=poly.fit_transform(X_test)
#train the model
model.fit(polyTR, Y_train)
#predict the values
PR_trainpred = model.predict(polyTR)
PR_testpred = model.predict(polyTE)
#print metrics for Polynomial Regression
PR_metrics_train = scores(Y_train,PR_trainpred)
PR_metrics_test = scores(Y_test, PR_testpred)
print('Train mse: ',PR_metrics_train[3])
print('Test mse: ',PR_metrics_test[3])
print('Train rmse: ',PR_metrics_train[0])
print('Test rmse: ',PR_metrics_test[0])
print('Train mae: ',PR_metrics_train[1])
print('Test mae: ',PR_metrics_test[1])
print('Train R2: ',PR_metrics_train[2])
print('Test R2: ',PR_metrics_test[2])
print()
#regression line for simple and polynomial
plt.scatter(X_test, Y_test, color='magenta')
plt.plot(X_test, PR_testpred, '*', color='darkgreen')
plt.show()
print()
#regression line for the predicted in test
plt.scatter(X_test, Y_test, color='magenta')
plt.plot(X_test, PR_testpred, '*', color='darkgreen')
plt.plot(xtest, SR_testpred, color='cyan', linewidth=2)
plt.show()
print()
#---------------------------------Multiple Linear Regression-----------------------
print('------------------Multiple Linear Regression: ')
print()
#linear model for how all features affect co2 emissions
targetcol = 6
selFeatures = list(df.columns.values)
del selFeatures[targetcol]
xtrainm,xtestm,ytrainm,ytestm = train_test_split(df[selFeatures],df['CO2EMISSIONS'],test_size = 0.3)
model.fit(xtrainm,ytrainm)
MR_trainpred = np.array(model.predict(xtrainm))
MR_testpred = np.array(model.predict(xtestm))
#print metrics for the simple linear regression model
MR_metrics_train = scores(ytrainm,MR_trainpred)
MR_metrics_test = scores(ytestm, MR_testpred)
print('Train mse: ',MR_metrics_train[3])
print('Test mse: ',MR_metrics_test[3])
print('Train rmse: ',MR_metrics_train[0])
print('Test rmse: ',MR_metrics_test[0])
print('Train mae: ',MR_metrics_train[1])
print('Test mae: ',MR_metrics_test[1])
print('Train R2: ',MR_metrics_train[2])
print('Test R2: ',MR_metrics_test[2])
print()
#-----------------------------------------Ridge-Lasso Simple Reg (engine size)
#Check diff with simple
print('------------------Ridge and Lasso Regression: ')
ridgeReg = Ridge(alpha=0.00001)
ridgeReg.fit(xtrain, ytrain)
lassoReg = Lasso(alpha=0.000001, max_iter = 10e5)
lassoReg.fit(xtrain, ytrain)
coeff_used = np.sum(lassoReg.coef_!=0)
RR_trainpred = ridgeReg.predict(xtrain)
RR_testpred = ridgeReg.predict(xtest)
LR_trainpred = lassoReg.predict(xtrain)
LR_testpred = lassoReg.predict(xtest)
RR_metrics_train = scores(ytrain,RR_trainpred)
RR_metrics_test = scores(ytest, RR_testpred)
LR_metrics_train = scores(ytrain,LR_trainpred)
LR_metrics_test = scores(ytest, LR_testpred)
#print Ridge
print('Train mse Ridge: ',RR_metrics_train[3])
print('Test mse Ridge: ',RR_metrics_test[3])
print('Train rmse Ridge: ',RR_metrics_train[0])
print('Test rmse Ridge: ',RR_metrics_test[0])
print('Train mae Ridge: ',RR_metrics_train[1])
print('Test mae Ridge: ',RR_metrics_test[1])
print('Train R2 Ridge: ',RR_metrics_train[2])
print('Test R2 Ridge: ',RR_metrics_test[2])
print()
#print Lasso
print('Train mse Lasso: ',LR_metrics_train[0])
print('Test mse Lasso: ',LR_metrics_test[0])
print('Train rmse Lasso: ',LR_metrics_train[1])
print('Test rmse Lasso: ',LR_metrics_test[1])
print('Train mae Lasso: ',LR_metrics_train[1])
print('Test mae Lasso: ',LR_metrics_test[1])
print('Train R2 Lasso: ',LR_metrics_train[2])
print('Test R2 Lasso: ',LR_metrics_test[2])
print('Number of features used: ', coeff_used)
print()
print()
#-----------------------------------------Ridge-Lasso Multiple Reg (all features)
#Check diff with multiple
print('------------------Ridge and Lasso Regression Multiple: ')
ridgeReg = Ridge(alpha=0.00001)
ridgeReg.fit(xtrainm, ytrainm)
lassoReg = Lasso(alpha=0.000001, max_iter = 10e5)
lassoReg.fit(xtrainm, ytrainm)
coeff_used2 = np.sum(lassoReg.coef_!=0)
RR_trainpred = ridgeReg.predict(xtrainm)
RR_testpred = ridgeReg.predict(xtestm)
LR_trainpred = lassoReg.predict(xtrainm)
LR_testpred = lassoReg.predict(xtestm)
RR_metrics_train = scores(ytrainm,RR_trainpred)
RR_metrics_test = scores(ytestm, RR_testpred)
LR_metrics_train = scores(ytrainm,LR_trainpred)
LR_metrics_test = scores(ytestm, LR_testpred)
#print Ridge
print('Train mse Ridge: ',RR_metrics_train[3])
print('Test mse Ridge: ',RR_metrics_test[3])
print('Train rmse Ridge: ',RR_metrics_train[0])
print('Test rmse Ridge: ',RR_metrics_test[0])
print('Train mae Ridge: ',RR_metrics_train[1])
print('Test mae Ridge: ',RR_metrics_test[1])
print('Train R2 Ridge: ',RR_metrics_train[2])
print('Test R2 Ridge: ',RR_metrics_test[2])
print()
#print Lasso
print('Train mse Lasso: ',LR_metrics_train[0])
print('Test mse Lasso: ',LR_metrics_test[0])
print('Train rmse Lasso: ',LR_metrics_train[1])
print('Test rmse Lasso: ',LR_metrics_test[1])
print('Train mae Lasso: ',LR_metrics_train[1])
print('Test mae Lasso: ',LR_metrics_test[1])
print('Train R2 Lasso: ',LR_metrics_train[2])
print('Test R2 Lasso: ',LR_metrics_test[2])
print('Number of features used: ', coeff_used2)
print()
print()