Kshape classifier

polars_ts.classification.kshape_classifier

k-Shape nearest-centroid classifier for time series.

KShapeClassifier

Nearest-centroid classifier using SBD and k-Shape centroids.

Fits k-Shape clustering per class on training data to extract representative centroids, then classifies test series by nearest centroid distance (SBD).

Parameters:

Name	Type	Description	Default
n_centroids_per_class	int	Number of k-Shape centroids per class. Default 1.	1
max_iter	int	Maximum k-Shape iterations per class. Default 100.	100

Examples:

>>> clf = KShapeClassifier(n_centroids_per_class=1)
>>> clf.fit(train_df, label_col="label")
>>> predictions = clf.predict(test_df)

fit(df, *, label_col='label')

Fit the classifier by computing k-Shape centroids for each class.

Parameters:

Name	Type	Description	Default
df	DataFrame	DataFrame with columns unique_id, y, and a label column.	required
label_col	str	Name of the column containing class labels.	'label'

Returns:

Type	Description
KShapeClassifier	self

predict(df)

Predict class labels for test time series.

Parameters:

Name	Type	Description	Default
df	DataFrame	DataFrame with columns unique_id and y.	required

Returns:

Type	Description
DataFrame	DataFrame with columns unique_id and predicted_label.