Decision Trees



In [1]:

    
from PIL import Image
import numpy as np
%matplotlib inline
import matplotlib
import matplotlib.pyplot as plt
from sklearn import datasets, tree
from sklearn import ensemble
matplotlib.style.use('bmh')
matplotlib.rcParams['figure.figsize']=(10,7)

MNIST



In [2]:

    
import gzip
import pickle
with gzip.open('../Week02/mnist.pkl.gz', 'rb') as f:
    train_set, validation_set, test_set = pickle.load(f, encoding='latin1')
    
train_X, train_y = train_set
test_X, test_y = test_set



In [3]:

    
train_X=train_X.reshape(-1, 28,28)[:,::2,::2].reshape(-1, 14*14)
test_X=test_X.reshape(-1, 28,28)[:,::2,::2].reshape(-1, 14*14)



In [4]:

    
clf = ensemble.AdaBoostClassifier(tree.DecisionTreeClassifier(max_depth=10), 
                                  n_estimators=50, 
                                  learning_rate=1,
                                 algorithm="SAMME.R")



In [5]:

    
%%timeit -n 1 -r 1
clf.fit(train_X, train_y)









    



1 loop, best of 1: 1min 13s per loop



In [6]:

    
%%timeit -n 1 -r 1
print(np.mean(clf.predict(train_X) == train_y))









    



0.99172
1 loop, best of 1: 1.2 s per loop



In [7]:

    
%%timeit -n 1 -r 1
print(np.mean(clf.predict(test_X) == test_y))









    



0.9346
1 loop, best of 1: 228 ms per loop



In [ ]: