def test_reproj_2():
pca = KPCA(n_components=2)
pca.fit(X1)
exp = pca.transform(X1[1, None])
assert_almost_equal(pca.X_projected_[1, None], exp, decimal=2)
def test_fail_array_transform():
pca = KPCA(n_components=2)
pca.fit(X1)
with pytest.raises(ValueError):
pca.transform(X1[1])
plt.xlabel('PC1')
plt.yticks([])
#plt.show()
plt.savefig('../figs/tutorial/mlxtendex1_4.png')
plt.close()
# Feature space linearly seperable at x=0, also data entirely 1-D, forming horizontal line.
#subspace like this can then be used as input in generalised classification models eg logistic regression.
import matplotlib.pyplot as plt
from sklearn.datasets import make_moons
#New data, two crescents, now using 100 samples per crescent,using random state seed of 5.
X2, y2 = make_moons(n_samples=200, random_state=5)
#Transform new dataset according to previous KPCA parameters.
X2_kpca = kpca.transform(X2)
#Initial data
plt.scatter(X_kpca[y==0, 0], X_kpca[y==0, 1],
color='red', marker='o', alpha=0.5, label='fit data')
plt.scatter(X_kpca[y==1, 0], X_kpca[y==1, 1],
color='blue', marker='^', alpha=0.5, label='fit data')
#New data projected onto new component axes
plt.scatter(X2_kpca[y2==0, 0], X2_kpca[y2==0, 1],
color='orange', marker='v',
alpha=0.2, label='new data')
plt.scatter(X2_kpca[y2==1, 0], X2_kpca[y2==1, 1],
color='cyan', marker='s',
alpha=0.2, label='new data')
def test_fail_array_transform():
pca = KPCA(n_components=2)
pca.fit(X1)
pca.transform(X1[1])
def test_reproj_2():
pca = KPCA(n_components=2)
pca.fit(X1)
exp = pca.transform(X1[1, None])
assert_almost_equal(pca.X_projected_[1, None], exp, decimal=2)
def test_fail_array_transform():
pca = KPCA(n_components=2)
pca.fit(X1)
exp = pca.transform(X1[1])
def test_fail_array_transform():
pca = KPCA(n_components=2)
pca.fit(X1)
with pytest.raises(ValueError):
pca.transform(X1[1])