test code & dataset for scikit-learn issue #365

README

code demonstrating the problem seen in issue #365

to run the example:
  tar -zxvf data.tgz
  python test.py

data.tgz

test.py

import numpy as np
from operator import itemgetter
from sklearn.feature_extraction.text import Vectorizer
from sklearn.neighbors import KNeighborsClassifier
from sklearn.datasets import load_files

data_train = load_files('data_train')
data_test = load_files('data_test')

categories = data_train.target_names

# split a training set and a test set
y_train, y_test = data_train.target, data_test.target

vectorizer = Vectorizer()

X_train = vectorizer.fit_transform(data_train.data)
X_test = vectorizer.transform(data_test.data)

vocabulary = np.array([t for t, i in sorted(vectorizer.vocabulary.iteritems(),
                                            key=itemgetter(1))])

knnfitted = KNeighborsClassifier(n_neighbors=1000,
                                 algorithm='brute').fit(X_train, y_train)
pred = knnfitted.predict(X_test)
print 1.0 * sum(pred) / len(pred)