Mercurial > hg > plosone_underreview
comparison tests/test_classification.py @ 58:d118b6ca8370 branch-tests
some changes in classification with random train/test split
author | Maria Panteli <m.x.panteli@gmail.com> |
---|---|
date | Thu, 21 Sep 2017 15:24:18 +0100 |
parents | e8084526f7e5 |
children | f9513664fe42 |
comparison
equal
deleted
inserted
replaced
57:dd86e49ae8bf | 58:d118b6ca8370 |
---|---|
17 X = np.random.randn(100, 3) | 17 X = np.random.randn(100, 3) |
18 # create 2 classes by shifting the entries of half the samples | 18 # create 2 classes by shifting the entries of half the samples |
19 X[-50:, :] = X[-50:, :] + 10 | 19 X[-50:, :] = X[-50:, :] + 10 |
20 Y = np.concatenate([np.repeat('a', 50), np.repeat('b', 50)]) | 20 Y = np.concatenate([np.repeat('a', 50), np.repeat('b', 50)]) |
21 X_train, X_test, Y_train, Y_test = train_test_split(X, Y, train_size=0.6, random_state=1, stratify=Y) | 21 X_train, X_test, Y_train, Y_test = train_test_split(X, Y, train_size=0.6, random_state=1, stratify=Y) |
22 accuracy, predictions = classification.confusion_matrix(X_train, Y_train, X_test, Y_test) | 22 accuracy, _ = classification.confusion_matrix(X_train, Y_train, X_test, Y_test) |
23 # expect perfect accuracy for this 'easy' dataset | 23 # expect perfect accuracy for this 'easy' dataset |
24 assert accuracy == 1.0 | 24 assert accuracy == 1.0 |
25 | 25 |