Source Code for Module parsimony.datasets.classification.dice5

 1  # -*- coding: utf-8 -*- 
 2  """ 
 3  Created on Tue Mar 18 10:56:04 2014 
 4   
 5  Copyright (c) 2013-2014, CEA/DSV/I2BM/Neurospin. All rights reserved. 
 6   
 7  @author:  Edouard Duchesnay 
 8  @email:   edouard.duchesnay@cea.fr 
 9  @license: BSD 3-clause. 
10  """ 
11  import numpy as np 
12  from ..regression import dice5 as dice5regression 
13   
14   
15  ############################################################################ 
16 -def load(n_samples=100, shape=(30, 30, 1), 
17           snr=2., sigma_logit=5., random_seed=None, 
18           **kwargs): 
19      """Generate classification samples (images + target variable) and beta. 
20      Call make_regression_struct then apply the logistic function: 
21      proba = 1. / (1 + exp(-(X * beta + noise)), then 
22      y = 1 if proba >= 0.5 else 0 
23   
24      Parameters 
25      ---------- 
26      See datasets.regression.dice5. 
27   
28      Returns 
29      ------- 
30      X3d : Numpy array of shape [n_sample, shape]. The input features. 
31   
32      y : Numpy array of shape [n_sample, 1]. The target variable. 
33   
34      beta3d : Float array of shape [shape]. It is the beta such that 
35              y = 1. / (1 + exp(-(X * beta + noise)). 
36   
37      proba : Numpy array of shape [n_sample, 1]. Samples posterior 
38              probabilities. 
39   
40      See also 
41      -------- 
42      regression.dice5.load() 
43   
44      Examples 
45      -------- 
46      >>> import numpy as np 
47      >>> np.random.seed(42) 
48      >>> import matplotlib.pyplot as plot 
49      >>> from  parsimony import datasets 
50      >>> X3d, y, beta3d, proba = datasets.classification.dice5.load( 
51      ...     n_samples=100, shape=(11, 11, 1), random_seed=1) 
52      """ 
53      X3d, y, beta3d = dice5regression.load(n_samples=n_samples, shape=shape, 
54                                            random_seed=random_seed, 
55                                            **kwargs) 
56      logit = y 
57      proba = np.reciprocal(1.0 + np.exp(-logit)) 
58      y = np.ones(y.shape) 
59      y[proba < 0.5] = 0 
60   
61      return X3d, y, beta3d, proba 
62