|
1
|
|
|
# Author: Simon Blanke |
|
2
|
|
|
# Email: [email protected] |
|
3
|
|
|
# License: MIT License |
|
4
|
|
|
|
|
5
|
|
|
import numpy as np |
|
6
|
|
|
import pandas as pd |
|
7
|
|
|
|
|
8
|
|
|
from sklearn.datasets import load_iris |
|
9
|
|
|
from sklearn.model_selection import cross_val_score |
|
10
|
|
|
from sklearn.tree import DecisionTreeClassifier |
|
11
|
|
|
from hyperactive import Hyperactive |
|
12
|
|
|
|
|
13
|
|
|
data = load_iris() |
|
14
|
|
|
X = data.data |
|
15
|
|
|
y = data.target |
|
16
|
|
|
|
|
17
|
|
|
|
|
18
|
|
|
def model(para, X, y): |
|
19
|
|
|
model = DecisionTreeClassifier( |
|
20
|
|
|
max_depth=para["max_depth"], |
|
21
|
|
|
min_samples_split=para["min_samples_split"], |
|
22
|
|
|
min_samples_leaf=para["min_samples_leaf"], |
|
23
|
|
|
) |
|
24
|
|
|
scores = cross_val_score(model, X, y, cv=3) |
|
25
|
|
|
|
|
26
|
|
|
return scores.mean() |
|
27
|
|
|
|
|
28
|
|
|
|
|
29
|
|
|
search_config = { |
|
30
|
|
|
model: { |
|
31
|
|
|
"max_depth": range(1, 21), |
|
32
|
|
|
"min_samples_split": range(2, 21), |
|
33
|
|
|
"min_samples_leaf": range(1, 21), |
|
34
|
|
|
} |
|
35
|
|
|
} |
|
36
|
|
|
|
|
37
|
|
|
warm_start = {model: {"max_depth": 2, "min_samples_split": 2, "min_samples_leaf": 2}} |
|
38
|
|
|
|
|
39
|
|
|
|
|
40
|
|
|
def test_func_return(): |
|
41
|
|
|
def model1(para, X, y): |
|
42
|
|
|
model = DecisionTreeClassifier( |
|
43
|
|
|
criterion=para["criterion"], |
|
44
|
|
|
max_depth=para["max_depth"], |
|
45
|
|
|
min_samples_split=para["min_samples_split"], |
|
46
|
|
|
min_samples_leaf=para["min_samples_leaf"], |
|
47
|
|
|
) |
|
48
|
|
|
scores = cross_val_score(model, X, y, cv=3) |
|
49
|
|
|
|
|
50
|
|
|
return scores.mean(), model |
|
51
|
|
|
|
|
52
|
|
|
search_config1 = { |
|
53
|
|
|
model1: { |
|
54
|
|
|
"criterion": ["gini", "entropy"], |
|
55
|
|
|
"max_depth": range(1, 21), |
|
56
|
|
|
"min_samples_split": range(2, 21), |
|
57
|
|
|
"min_samples_leaf": range(1, 21), |
|
58
|
|
|
} |
|
59
|
|
|
} |
|
60
|
|
|
|
|
61
|
|
|
opt = Hyperactive(X, y) |
|
62
|
|
|
opt.search(search_config1) |
|
63
|
|
|
|
|
64
|
|
|
|
|
65
|
|
|
def test_n_jobs_2(): |
|
66
|
|
|
opt = Hyperactive(X, y) |
|
67
|
|
|
opt.search(search_config, n_jobs=2) |
|
68
|
|
|
|
|
69
|
|
|
|
|
70
|
|
|
def test_n_jobs_4(): |
|
71
|
|
|
opt = Hyperactive(X, y) |
|
72
|
|
|
opt.search(search_config, n_jobs=4) |
|
73
|
|
|
|
|
74
|
|
|
|
|
75
|
|
|
def test_positional_args(): |
|
76
|
|
|
opt0 = Hyperactive(X, y, random_state=False) |
|
77
|
|
|
opt0.search(search_config) |
|
78
|
|
|
|
|
79
|
|
|
opt1 = Hyperactive(X, y, random_state=1) |
|
80
|
|
|
opt1.search(search_config) |
|
81
|
|
|
|
|
82
|
|
|
opt2 = Hyperactive(X, y, random_state=1) |
|
83
|
|
|
opt2.search(search_config) |
|
84
|
|
|
|
|
85
|
|
|
|
|
86
|
|
|
def test_random_state(): |
|
87
|
|
|
opt0 = Hyperactive(X, y, random_state=False) |
|
88
|
|
|
opt0.search(search_config) |
|
89
|
|
|
|
|
90
|
|
|
opt1 = Hyperactive(X, y, random_state=0) |
|
91
|
|
|
opt1.search(search_config) |
|
92
|
|
|
|
|
93
|
|
|
opt2 = Hyperactive(X, y, random_state=1) |
|
94
|
|
|
opt2.search(search_config) |
|
95
|
|
|
|
|
96
|
|
|
|
|
97
|
|
|
def test_max_time(): |
|
98
|
|
|
opt0 = Hyperactive(X, y) |
|
99
|
|
|
opt0.search(search_config, max_time=0.001) |
|
100
|
|
|
|
|
101
|
|
|
|
|
102
|
|
|
def test_memory(): |
|
103
|
|
|
opt0 = Hyperactive(X, y, memory=True) |
|
104
|
|
|
opt0.search(search_config) |
|
105
|
|
|
|
|
106
|
|
|
opt1 = Hyperactive(X, y, memory=False) |
|
107
|
|
|
opt1.search(search_config) |
|
108
|
|
|
|
|
109
|
|
|
|
|
110
|
|
|
def test_verbosity(): |
|
111
|
|
|
opt0 = Hyperactive(X, y, verbosity=0) |
|
112
|
|
|
opt0.search(search_config) |
|
113
|
|
|
|
|
114
|
|
|
opt0 = Hyperactive(X, y, verbosity=0) |
|
115
|
|
|
opt0.search(search_config, n_jobs=2) |
|
116
|
|
|
|
|
117
|
|
|
opt1 = Hyperactive(X, y, verbosity=1) |
|
118
|
|
|
opt1.search(search_config) |
|
119
|
|
|
|
|
120
|
|
|
opt0 = Hyperactive(X, y, verbosity=1) |
|
121
|
|
|
opt0.search(search_config) |
|
122
|
|
|
|
|
123
|
|
|
opt1 = Hyperactive(X, y, verbosity=2) |
|
124
|
|
|
opt1.search(search_config, n_jobs=2) |
|
125
|
|
|
|
|
126
|
|
|
|
|
127
|
|
|
def test_scatter_init(): |
|
128
|
|
|
opt = Hyperactive(X, y) |
|
129
|
|
|
opt.search(search_config, scatter_init=10) |
|
130
|
|
|
|
|
131
|
|
|
|
|
132
|
|
|
def test_optimizer_args(): |
|
133
|
|
|
opt = Hyperactive(X, y) |
|
134
|
|
|
opt.search(search_config, optimizer={"HillClimbing": {"epsilon": 0.1}}) |
|
135
|
|
|
|
|
136
|
|
|
|
|
137
|
|
|
def test_scatter_init_and_warm_start(): |
|
138
|
|
|
opt = Hyperactive(X, y) |
|
139
|
|
|
opt.search(search_config, warm_start=warm_start, scatter_init=10) |
|
140
|
|
|
|
|
141
|
|
|
opt = Hyperactive(X, y) |
|
142
|
|
|
opt.search(search_config, warm_start=warm_start, scatter_init=10) |
|
143
|
|
|
|
|
144
|
|
|
|
|
145
|
|
|
def test_warm_start_multiple_jobs(): |
|
146
|
|
|
opt = Hyperactive(X, y) |
|
147
|
|
|
opt.search(search_config, n_jobs=4, warm_start=warm_start) |
|
148
|
|
|
|
|
149
|
|
|
|
|
150
|
|
|
def test_warm_start(): |
|
151
|
|
|
opt = Hyperactive(X, y) |
|
152
|
|
|
opt.search(search_config, n_jobs=1, warm_start=warm_start) |
|
153
|
|
|
|
|
154
|
|
|
|
|
155
|
|
|
def test_get_search_path(): |
|
156
|
|
|
opt = Hyperactive(X, y, get_search_path=True) |
|
157
|
|
|
opt.search(search_config) |
|
158
|
|
|
|
|
159
|
|
|
opt = Hyperactive(X, y, optimizer="ParticleSwarm", get_search_path=True) |
|
160
|
|
|
opt.search(search_config) |
|
161
|
|
|
|
SimonBlanke /
Hyperactive
