def test_fit_transform(params):
"""Check that fit and transform and fit_transform yield same results."""
transformer = WEASELMUSE(**params)
arr_1 = transformer.fit(X, y).transform(X)
arr_2 = transformer.fit_transform(X, y)
if transformer.sparse:
assert (arr_1 != arr_2).nnz == 0
else:
np.testing.assert_array_equal(arr_1, arr_2)
def test_n_estimators():
"""Check that the number of estimators is the number of features."""
transformer = WEASELMUSE().fit(X, y)
assert len(transformer._estimators) == n_features
assert len(transformer._estimators_diff) == n_features
def test_output_dtype(params, type_desired):
"""Check that the output dtype is the expected one."""
transformer = WEASELMUSE(**params)
output = transformer.fit_transform(X, y)
assert isinstance(output, type_desired)
def test_output_ndim(params):
"""Check that the number of dimensions is always 2."""
transformer = WEASELMUSE(**params)
ndim_actual = transformer.fit_transform(X, y).ndim
assert ndim_actual == 2
# License: BSD-3-Clause
import numpy as np
import matplotlib.pyplot as plt
from pyts.datasets import load_basic_motions
from pyts.multivariate.transformation import WEASELMUSE
from sklearn.preprocessing import LabelEncoder
# Toy dataset
X_train, _, y_train, _ = load_basic_motions(return_X_y=True)
y_train = LabelEncoder().fit_transform(y_train)
# WEASEL+MUSE transformation
transformer = WEASELMUSE(word_size=2,
n_bins=2,
window_sizes=[12, 36],
chi2_threshold=15,
sparse=False)
X_weasel = transformer.fit_transform(X_train, y_train)
# Visualize the transformation for the first time series
plt.figure(figsize=(6, 4))
vocabulary_length = len(transformer.vocabulary_)
width = 0.3
plt.bar(np.arange(vocabulary_length) - width / 2,
X_weasel[y_train == 0][0],
width=width,
label='First time series in class 0')
plt.bar(np.arange(vocabulary_length) + width / 2,
X_weasel[y_train == 1][0],
width=width,