Shapelet forests and extremly randomized shapelet trees

In this example, we explore the training time and predictive performance of the random shapelet forest and the extremly randomized shapelet trees algorithm.

import numpy as np
from sklearn.model_selection import cross_validate

from wildboar.datasets import load_dataset
from wildboar.ensemble import ExtraShapeletTreesClassifier, ShapeletForestClassifier

random_state = 1234

First, we load the datasets merging any existing training and testing partitions.

x, y = load_dataset("Beef")

Next, we setup the two classifiers we want to compare.

classifiers = {
    "Shapelet forest": ShapeletForestClassifier(
        n_shapelets=10,
        metric="scaled_euclidean",
        n_jobs=-1,
        random_state=random_state,
    ),
    "Extra Shapelet Trees": ExtraShapeletTreesClassifier(
        metric="scaled_euclidean",
        n_jobs=-1,
        random_state=random_state,
    ),
}

Finally, we iterate over the classifiers and compute the cross-validation area under ROC. We also print the time it takes to train the algorithms and their respective performance.

for name, clf in classifiers.items():
    score = cross_validate(clf, x, y, scoring="roc_auc_ovo", n_jobs=1)
    print(f"Classifier: {name}")
    print(" - fit-time:   %.2f" % np.mean(score["fit_time"]))
    print(" - test-score: %.2f" % np.mean(score["test_score"]))

Classifier: Shapelet forest
 - fit-time:   0.75
 - test-score: 0.88
Classifier: Extra Shapelet Trees
 - fit-time:   0.17
 - test-score: 0.87