sklearn knn与kmeans

《机器学习技法》最后一次作业，共有三个编程实验，倒数两个是knn和kmeans。照例用sklearn做，很快。

knn：分别求k=1和k=5时的Ein和Eout.

from sklearn import neighbors
import numpy as np

train_data = np.loadtxt('hw4_knn_train.dat')
train_x = train_data[:, :-1]
train_y = train_data[:,-1]

test_data = np.loadtxt('hw4_knn_test.dat')
test_x = test_data[:, :-1]
test_y = test_data[:,-1]

clf = neighbors.KNeighborsClassifier(n_neighbors=5)
clf.fit(train_x, train_y)

err_in = 1 - clf.score(train_x, train_y)
err_out = 1 - clf.score(test_x, test_y)
print err_in, err_out

kmeans：

分别求k=2和k=10时的Ein

from __future__ import division
from sklearn.cluster import KMeans
import numpy as np

data = np.loadtxt('hw4_kmeans_train.dat')
N = data.shape[0]

repeat = 500
e_in = 0

for i in range(repeat):
    clf = KMeans(n_clusters=10, init='random', max_iter=300)
    clf.fit(data)
    e_in = e_in + clf.inertia_/N
    
print e_in/repeat