Hands-On Ensemble Learning with Python

Extra trees for classification

Apart from conventional Random Forests, scikit-learn also implements Extra Trees. The classification implementation lies in the ExtraTreesClassifier, in the sklearn.ensemble package. Here, we repeat the hand-written digit recognition example, using the Extra Trees classifier:

# --- SECTION 1 ---# Libraries and data loadingfrom sklearn.datasets import load_digitsfrom sklearn.ensemble import ExtraTreesClassifierfrom sklearn import metricsimport numpy as npdigits = load_digits()train_size = 1500train_x, train_y = digits.data[:train_size], digits.target[:train_size]test_x, test_y = digits.data[train_size:], digits.target[train_size:]np.random.seed(123456)# --- SECTION 2 ---# Create the ensembleensemble_size = 500 ...