Explore Library
Code QuizIntermediate

Additive Prediction Accumulation

Finding why only the last tree affects the boosted prediction.

Codepython
import numpy as np

def predict_ensemble(trees, X, init, learning_rate):
    pred = np.full(len(X), init)
    for tree in trees:
        # accumulate each tree's contribution
        pred = learning_rate * tree.predict(X)
    return pred

What is the bug in this ensemble prediction loop?