|
1
|
|
|
# Author: Simon Blanke |
|
2
|
|
|
# Email: [email protected] |
|
3
|
|
|
# License: MIT License |
|
4
|
|
|
|
|
5
|
|
|
|
|
6
|
|
|
import numpy as np |
|
7
|
|
|
|
|
8
|
|
|
from sklearn.gaussian_process import GaussianProcessRegressor |
|
9
|
|
|
from sklearn.gaussian_process.kernels import Matern |
|
10
|
|
|
from sklearn.ensemble import ExtraTreesRegressor as _ExtraTreesRegressor_ |
|
11
|
|
|
from sklearn.ensemble import ExtraTreesRegressor as _RandomForestRegressor_ |
|
12
|
|
|
|
|
13
|
|
|
|
|
14
|
|
|
def _return_std(X, trees, predictions, min_variance): |
|
15
|
|
|
std = np.zeros(len(X)) |
|
16
|
|
|
|
|
17
|
|
|
for tree in trees: |
|
18
|
|
|
var_tree = tree.tree_.impurity[tree.apply(X)] |
|
19
|
|
|
var_tree[var_tree < min_variance] = min_variance |
|
20
|
|
|
mean_tree = tree.predict(X) |
|
21
|
|
|
std += var_tree + mean_tree ** 2 |
|
22
|
|
|
|
|
23
|
|
|
std /= len(trees) |
|
24
|
|
|
std -= predictions ** 2.0 |
|
25
|
|
|
std[std < 0.0] = 0.0 |
|
26
|
|
|
std = std ** 0.5 |
|
27
|
|
|
# print("std", std) |
|
28
|
|
|
return std |
|
29
|
|
|
|
|
30
|
|
|
|
|
31
|
|
|
class TreeEnsembleBase: |
|
32
|
|
|
def __init__(self, min_variance=0.0, **kwargs): |
|
33
|
|
|
self.min_variance = min_variance |
|
34
|
|
|
super().__init__(**kwargs) |
|
35
|
|
|
|
|
36
|
|
|
def fit(self, X, y): |
|
37
|
|
|
super().fit(X, np.ravel(y)) |
|
38
|
|
|
|
|
39
|
|
|
def predict(self, X, return_std=False): |
|
40
|
|
|
mean = super().predict(X) |
|
41
|
|
|
|
|
42
|
|
|
if return_std: |
|
43
|
|
|
if self.criterion != "mse": |
|
44
|
|
|
raise ValueError( |
|
45
|
|
|
"Expected impurity to be 'mse', got %s instead" % self.criterion |
|
46
|
|
|
) |
|
47
|
|
|
std = _return_std(X, self.estimators_, mean, self.min_variance) |
|
48
|
|
|
return mean.reshape(-1, 1), std |
|
49
|
|
|
return mean.reshape(-1, 1) |
|
50
|
|
|
|
|
51
|
|
|
|
|
52
|
|
|
class RandomForestRegressor(TreeEnsembleBase, _RandomForestRegressor_): |
|
53
|
|
|
def __init__(self, min_variance=0.0, **kwargs): |
|
54
|
|
|
super().__init__(**kwargs) |
|
55
|
|
|
|
|
56
|
|
|
|
|
57
|
|
|
class ExtraTreesRegressor(TreeEnsembleBase, _ExtraTreesRegressor_): |
|
58
|
|
|
def __init__(self, min_variance=0.0, **kwargs): |
|
59
|
|
|
super().__init__(**kwargs) |
|
60
|
|
|
|
|
61
|
|
|
|
|
62
|
|
|
class GPR: |
|
63
|
|
|
def __init__(self): |
|
64
|
|
|
self.gpr = GaussianProcessRegressor(kernel=Matern(nu=2.5), normalize_y=True) |
|
65
|
|
|
|
|
66
|
|
|
def fit(self, X, y): |
|
67
|
|
|
self.gpr.fit(X, y) |
|
68
|
|
|
|
|
69
|
|
|
def predict(self, X, return_std=False): |
|
70
|
|
|
return self.gpr.predict(X, return_std=return_std) |
|
71
|
|
|
|
SimonBlanke /
Hyperactive
