In [1]:

    

import sklearn
import matplotlib.pyplot as plt


    

In [2]:

    

from sklearn import datasets


    

In [3]:

    

from sklearn.mixture import GaussianMixture


    

In [4]:

    

import numpy


    

In [5]:

    

iris = datasets.load_iris()


    

In [6]:

    

from sklearn.model_selection import StratifiedKFold


    

In [7]:

    

skf = StratifiedKFold(n_splits=4)


    

In [8]:

    

train_index, test_index = next(iter(skf.split(iris.data, iris.target)))


    

In [23]:

    

X_train = iris.data[train_index]
y_train = iris.target[train_index]
X_test = iris.data[test_index]
y_test = iris.target[test_index]


    

In [10]:

    

n_classes = len(numpy.unique(y_train))


    

In [11]:

    

n_classes


    

    
Out[11]:





3

In [14]:

    

estimators = dict((cov_type, GaussianMixture(n_components=n_classes, covariance_type=cov_type, max_iter=20, random_state=0))  for cov_type in ['spherical', 'diag', 'tied', 'full'])


    

In [16]:

    

y_train


    

    
Out[16]:





array([0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
       0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1,
       1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
       1, 1, 1, 1, 1, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2,
       2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2])

In [17]:

    

estd = estimators['diag']


    

In [26]:

    

estd.means_init = numpy.array([X_train[y_train == i].mean(axis=0) for i in range(n_classes)])


    

In [27]:

    

estd.means_init


    

    
Out[27]:





array([[ 5.04594595,  3.45135135,  1.46486486,  0.25675676],
       [ 5.91621622,  2.77297297,  4.25945946,  1.32432432],
       [ 6.59459459,  2.98378378,  5.51351351,  2.02432432]])

In [ ]: