-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathmodel.py
195 lines (177 loc) · 7.59 KB
/
model.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
import pandas as pd
import numpy as np
import sklearn.metrics as metric
from sklearn.linear_model import LinearRegression, LassoLars, TweedieRegressor
from sklearn.preprocessing import PolynomialFeatures
def regression_models(X_train, y_train, X_validate, y_validate):
'''
Takes in X_train, y_train, X_validate, y_validate and runs
different models and produces df with RMSE and r^2 scores
for each model on train and validate.
'''
train_predictions = pd.DataFrame(y_train)
validate_predictions = pd.DataFrame(y_validate)
# create the metric_df as a blank dataframe
metric_df = pd.DataFrame()
#OLS Model
lm = LinearRegression(normalize=True)
lm.fit(X_train, y_train)
train_predictions['lm'] = lm.predict(X_train)
# predict validate
validate_predictions['lm'] = lm.predict(X_validate)
metric_df = make_metric_df(y_train, train_predictions.lm, y_validate, validate_predictions.lm, metric_df, model_name = 'OLS Regressor')
#Lasso Lars
# create the model object
lars = LassoLars(alpha=1)
lars.fit(X_train, y_train)
# predict train
train_predictions['lars'] = lars.predict(X_train)
# predict validate
validate_predictions['lars'] = lars.predict(X_validate)
metric_df = make_metric_df(y_train, train_predictions.lars, y_validate, validate_predictions.lars, metric_df, model_name = 'Lasso_alpha_1')
#Tweedie Regressor/GLM
# create the model object
glm = TweedieRegressor(power=1, alpha=0)
glm.fit(X_train, y_train)
# predict train
train_predictions['glm'] = glm.predict(X_train)
# predict validate
validate_predictions['glm'] = glm.predict(X_validate)
metric_df = make_metric_df(y_train, train_predictions.glm, y_validate, validate_predictions.glm, metric_df, model_name = 'GLM')
# make the polynomial features to get a new set of features
pf = PolynomialFeatures(degree=2)
# fit and transform X_train_scaled
X_train_degree2 = pf.fit_transform(X_train)
# transform X_validate_scaled & X_test_scaled
X_validate_degree2 = pf.transform(X_validate)
# create the model object
lm2 = LinearRegression(normalize=True)
lm2.fit(X_train_degree2, y_train)
# predict train
train_predictions['poly_2'] = lm2.predict(X_train_degree2)
# predict validate
validate_predictions['poly_2'] = lm2.predict(X_validate_degree2)
metric_df = make_metric_df(y_train, train_predictions.poly_2, y_validate, validate_predictions.poly_2, metric_df, model_name = 'Quadratic')
return metric_df
def make_metric_df(y_train, y_train_pred, y_validate, y_validate_pred, metric_df,model_name ):
'''
Takes in y_train, y_train_pred, y_validate, y_validate_pred, and a df
returns a df of RMSE and r^2 score for the model on train and validate
'''
if metric_df.size ==0:
metric_df = pd.DataFrame(data=[
{
'model': model_name,
f'RMSE_train': metric.mean_squared_error(
y_train,
y_train_pred) ** .5,
f'r^2_train': metric.explained_variance_score(
y_train,
y_train_pred),
f'RMSE_validate': metric.mean_squared_error(
y_validate,
y_validate_pred) ** .5,
f'r^2_validate': metric.explained_variance_score(
y_validate,
y_validate_pred)
}])
return metric_df
else:
return metric_df.append(
{
'model': model_name,
f'RMSE_train': metric.mean_squared_error(
y_train,
y_train_pred) ** .5,
f'r^2_train': metric.explained_variance_score(
y_train,
y_train_pred),
f'RMSE_validate': metric.mean_squared_error(
y_validate,
y_validate_pred) ** .5,
f'r^2_validate': metric.explained_variance_score(
y_validate,
y_validate_pred)
}, ignore_index=True)
def baseline_models(y_train, y_validate):
'''
Takes in y_train and y_validate and returns a df of
baseline_mean and baseline_median and how they perform
'''
train_predictions = pd.DataFrame(y_train)
validate_predictions = pd.DataFrame(y_validate)
y_pred_mean = y_train.mean()
train_predictions['y_pred_mean'] = y_pred_mean
validate_predictions['y_pred_mean'] = y_pred_mean
y_pred_median = y_train.median()
train_predictions['y_pred_median'] = y_pred_median
validate_predictions['y_pred_median'] = y_pred_median
# create the metric_df as a blank dataframe
metric_df = pd.DataFrame(data=[
{
'model': 'mean_baseline',
'RMSE_train': metric.mean_squared_error(
y_train,
train_predictions['y_pred_mean']) ** .5,
'RMSE_validate': metric.mean_squared_error(
y_validate,
validate_predictions['y_pred_mean']) ** .5,
'Difference': (( metric.mean_squared_error(
y_train,
train_predictions['y_pred_mean']) ** .5)-(metric.mean_squared_error(
y_validate,
validate_predictions['y_pred_mean']) ** .5))
}])
return metric_df.append(
{
'model': 'median_baseline',
'RMSE_train': metric.mean_squared_error(
y_train,
train_predictions['y_pred_median']) ** .5,
'RMSE_validate': metric.mean_squared_error(
y_validate,
validate_predictions['y_pred_median']) ** .5,
'Difference': (( metric.mean_squared_error(
y_train,
train_predictions['y_pred_median']) ** .5)-(metric.mean_squared_error(
y_validate,
validate_predictions['y_pred_median']) ** .5))
}, ignore_index=True)
def best_model(X_train, y_train, X_validate, y_validate, X_test, y_test):
'''
Takes in X_train, y_train, X_validate, y_validate, X_test, y_test
and returns a df with the RMSE and r^2 score on train, validate and test
'''
# make the polynomial features to get a new set of features
pf = PolynomialFeatures(degree=2)
# fit and transform X_train_scaled
X_train_degree2 = pf.fit_transform(X_train)
# transform X_validate_scaled & X_test_scaled
X_validate_degree2 = pf.transform(X_validate)
X_test_degree2 = pf.transform(X_test)
# create the model object
lm2 = LinearRegression(normalize=True)
lm2.fit(X_train_degree2, y_train)
metric_df = pd.DataFrame(data=[
{
'model': 'Quadratic',
f'RMSE_train': metric.mean_squared_error(
y_train,
lm2.predict(X_train_degree2)) ** .5,
f'r^2_train': metric.explained_variance_score(
y_train,
lm2.predict(X_train_degree2)),
f'RMSE_validate': metric.mean_squared_error(
y_validate,
lm2.predict(X_validate_degree2)) ** .5,
f'r^2_validate': metric.explained_variance_score(
y_validate,
lm2.predict(X_validate_degree2)),
f'RMSE_test': metric.mean_squared_error(
y_test,
lm2.predict(X_test_degree2)) ** .5,
f'r^2_test': metric.explained_variance_score(
y_test,
lm2.predict(X_test_degree2))
}])
return metric_df