-
Notifications
You must be signed in to change notification settings - Fork 1
/
cross_validate.py
163 lines (122 loc) · 6.53 KB
/
cross_validate.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
#!/usr/bin/env python
"""
Cross-validation with a few classifiers.
"""
import pandas as pd
import numpy as np
from time import clock
from itertools import product
### Load and prepare data
train_file = 'dataset/numerai_training_data.csv'
start = clock()
train_frame = pd.read_csv(train_file)
print('Loaded {:d} train entries in {:.0f} seconds.'.format(
len(train_frame), clock() - start))
# Remove validation column, not used here
train_frame.drop('validation', axis = 1 , inplace = True)
# Separate train data and label
label = train_frame['target']
train_frame.drop('target', axis = 1, inplace = True)
# One-hot encode of categorical variable
# Encode column in train, then drop original column
train_dummies = pd.get_dummies(train_frame['c1'])
train = pd.concat((train_frame.drop('c1', axis = 1), train_dummies.astype(int)), axis = 1)
### Select classifiers
clfs = []
from sklearn.ensemble import RandomForestClassifier as RF
rf1 = RF(n_estimators = 10, verbose = True)
rf2 = RF(n_estimators = 100, verbose = True)
rf3 = RF(n_estimators = 1000, verbose = True)
rf4 = RF(n_estimators = 10000, verbose = True)
from sklearn.linear_model import SGDClassifier
sgd = SGDClassifier()
from sklearn.svm import LinearSVC
lsvc = LinearSVC(tol = 0.01, C = 1)
from sklearn.ensemble import ExtraTreesClassifier
etc2 = ExtraTreesClassifier(n_estimators = 100, max_depth = None, min_samples_split = 1, random_state = 0)
etc3 = ExtraTreesClassifier(n_estimators = 1000, max_depth = None, min_samples_split = 1, random_state = 0)
# Logistic regression with preprocessor
from sklearn.pipeline import make_pipeline
from sklearn.preprocessing import Normalizer, MinMaxScaler, StandardScaler
from sklearn.linear_model import LogisticRegression as LR
lr = LR()
transformers = [ MinMaxScaler(), StandardScaler(),
Normalizer( norm = 'l1' ), Normalizer( norm = 'l2' ) ]
clfs += [make_pipeline(p) for p in product(transformers, [LR()])]
# Classifiers from Scikit Flow
# Optimizer choices: SGD, Adam, Adagrad
from skflow import TensorFlowLinearClassifier
tflc = TensorFlowLinearClassifier(n_classes = 1, batch_size = 256, steps = 1400, learning_rate = 0.01, optimizer = 'Adagrad')
from skflow import TensorFlowLinearRegressor
tflr = TensorFlowLinearRegressor(n_classes = 1, batch_size = 256, steps = 1400, learning_rate = 0.01, optimizer = 'Adagrad')
from skflow import TensorFlowDNNClassifier
tfdnnc = TensorFlowDNNClassifier(hidden_units = [100, 200, 200, 200, 100],
n_classes = 1, batch_size = 256, steps = 1000, learning_rate = 0.01, optimizer = 'Adagrad')
clfs += [lr, lsvc, sgd, rf1, rf2, rf3, rf4, etc2, etc3]
# clfs += [tflc, tflr, tfdnnc]
### Cross validation
from sklearn.cross_validation import cross_val_score
for clf in clfs:
print(clf)
start = clock()
scores = cross_val_score(clf, train, label, scoring = 'roc_auc', cv = 10, verbose = 1)
print(
"Performed {:d}-fold cross validation in {:.0f} seconds with ROC AUC: mean {:0.4f} std {:0.4f}.".format(
len(scores), clock() - start, scores.mean(), scores.std() ))
"""
Results
LogisticRegression(C=1.0, class_weight=None, dual=False, fit_intercept=True,
intercept_scaling=1, penalty=l2, random_state=None, tol=0.0001)
Performed 10-fold cross validation in 3 seconds with ROC AUC: mean 0.5254 std 0.0044.
LinearSVC(C=1, class_weight=None, dual=True, fit_intercept=True,
intercept_scaling=1, loss=l2, multi_class=ovr, penalty=l2,
random_state=None, tol=0.01, verbose=0)
Performed 10-fold cross validation in 116 seconds with ROC AUC: mean 0.5051 std 0.0186.
SGDClassifier(alpha=0.0001, class_weight=None, epsilon=0.1, eta0=0.0,
fit_intercept=True, l1_ratio=0.15, learning_rate=optimal,
loss=hinge, n_iter=5, n_jobs=1, penalty=l2, power_t=0.5,
random_state=None, rho=None, shuffle=False, verbose=0,
warm_start=False)
Performed 10-fold cross validation in 1 seconds with ROC AUC: mean 0.5002 std 0.0135.
RandomForestClassifier(bootstrap=True, compute_importances=None,
criterion=gini, max_depth=None, max_features=auto,
min_density=None, min_samples_leaf=1, min_samples_split=2,
n_estimators=10, n_jobs=1, oob_score=False, random_state=None,
verbose=True)
Performed 10-fold cross validation in 35 seconds with ROC AUC: mean 0.5057 std 0.0058.
RandomForestClassifier(bootstrap=True, compute_importances=None,
criterion=gini, max_depth=None, max_features=auto,
min_density=None, min_samples_leaf=1, min_samples_split=2,
n_estimators=100, n_jobs=1, oob_score=False, random_state=None,
verbose=True)
Performed 10-fold cross validation in 351 seconds with ROC AUC: mean 0.5213 std 0.0082.
RandomForestClassifier(bootstrap=True, compute_importances=None,
criterion=gini, max_depth=None, max_features=auto,
min_density=None, min_samples_leaf=1, min_samples_split=2,
n_estimators=1000, n_jobs=1, oob_score=False,
random_state=None, verbose=True)
Performed 10-fold cross validation in 3307 seconds with ROC AUC: mean 0.5279 std 0.0054.
ExtraTreesClassifier(bootstrap=False, compute_importances=None,
criterion=gini, max_depth=None, max_features=auto,
min_density=None, min_samples_leaf=1, min_samples_split=1,
n_estimators=100, n_jobs=1, oob_score=False, random_state=0,
verbose=0)
Performed 10-fold cross validation in 143 seconds with ROC AUC: mean 0.5228 std 0.0077.
ExtraTreesClassifier(bootstrap=False, compute_importances=None,
criterion=gini, max_depth=None, max_features=auto,
min_density=None, min_samples_leaf=1, min_samples_split=1,
n_estimators=1000, n_jobs=1, oob_score=False, random_state=0,
verbose=0)
Performed 10-fold cross validation in 1433 seconds with ROC AUC: mean 0.5255 std 0.0067.
LogisticRegression(C=1.0, class_weight=None, dual=False, fit_intercept=True,
intercept_scaling=1, penalty=l2, random_state=None, tol=0.0001)
Performed 10-fold cross validation in 3 seconds with ROC AUC: mean 0.5254 std 0.0044.
MinMaxScaler(copy=True, feature_range=(0, 1))
Performed 10-fold cross validation in 4 seconds with ROC AUC: mean 0.5354 std 0.0062.
StandardScaler(copy=True, with_mean=True, with_std=True)
Performed 10-fold cross validation in 5 seconds with ROC AUC: mean 0.5354 std 0.0062.
Normalizer(copy=True, norm=l1)
Performed 10-fold cross validation in 3 seconds with ROC AUC: mean 0.5254 std 0.0049.
Normalizer(copy=True, norm=l2)
Performed 10-fold cross validation in 3 seconds with ROC AUC: mean 0.5261 std 0.0051.
"""