vary_sample_size.py

import time
from sklearn.datasets import load_svmlight_file
from sklearn.svm import LinearSVC
from sklearn.neural_network import MLPClassifier
from sklearn.metrics import accuracy_score

# =========================
# Dataset configuration
# =========================
DATASETS = [
    ("a1a", 1605, 30956),
    ("a2a", 2265, 30296),
    ("a3a", 3185, 29376),
    ("a4a", 4781, 27780),
    ("a5a", 6414, 26147),
    ("a6a", 11220, 21341),
    ("a7a", 16100, 16461),
    ("a8a", 22696, 9865),
    ("a9a", 32561, 16281),
]

FEATURE_DIM = 123
TRAIN_DIR = "dataset/train"
TEST_DIR = "dataset/test"

# =========================
# Models
# =========================
def build_svm():
    return LinearSVC(
        C=1.0,
        max_iter=5000
    )

def build_mlp():
    return MLPClassifier(
        hidden_layer_sizes=(50,),
        activation="relu",
        alpha=1e-4,
        solver="adam",
        max_iter=50,
        random_state=42
    )

# =========================
# Load dataset
# =========================
def load_dataset(name):
    X_train, y_train = load_svmlight_file(
        f"{TRAIN_DIR}/{name}", n_features=FEATURE_DIM
    )
    X_test, y_test = load_svmlight_file(
        f"{TEST_DIR}/{name}.t", n_features=FEATURE_DIM
    )
    return X_train, y_train, X_test, y_test

# =========================
# Run experiment
# =========================
def run_experiment(model, X_train, y_train, X_test, y_test):
    start = time.time()
    model.fit(X_train, y_train)
    train_time = time.time() - start

    y_pred = model.predict(X_test)
    acc = accuracy_score(y_test, y_pred)
    return acc, train_time

# =========================
# Main
# =========================
def main():
    print("| Dataset | SVM Acc | MLP Acc | SVM Time (s) | MLP Time (s) |")
    print("|---------|---------|---------|--------------|--------------|")

    for name, train_size, test_size in DATASETS:
        X_train, y_train, X_test, y_test = load_dataset(name)

        # SVM
        svm = build_svm()
        svm_acc, svm_time = run_experiment(
            svm, X_train, y_train, X_test, y_test
        )

        # MLP
        mlp = build_mlp()
        mlp_acc, mlp_time = run_experiment(
            mlp, X_train, y_train, X_test, y_test
        )

        print(
            f"| {name} | "
            f"{svm_acc:.4f} | "
            f"{mlp_acc:.4f} | "
            f"{svm_time:.2f} | "
            f"{mlp_time:.2f} |"
        )

if __name__ == "__main__":
    main()