Squash, Then Classify

PCA is a transformer like a scaler, so it can sit inside a pipeline in front of a model. How few directions can a classifier get away with?

Task: write accuracy_with_directions(n_components).

• Load load_breast_cancer(return_X_y=True).
• Split with train_test_split(X, y, test_size=0.25, random_state=0, stratify=y).
• Build make_pipeline(StandardScaler(), PCA(n_components=n_components), LogisticRegression(max_iter=1000)) and fit it on the training rows.
• Return its accuracy on the test rows, as a plain float rounded to 3 decimal places.