PriyaPbs/ModelOptix

1import pandas as pd2from sklearn.feature_selection import VarianceThreshold3 4def select_features(X, y, preference):5    selector = VarianceThreshold(threshold=0.01)6    X_var = selector.fit_transform(X)7 8    df = pd.DataFrame(X_var)9 10    corr = df.corrwith(pd.Series(y)).abs()11    corr = corr.sort_values(ascending=False)12 13    if preference == "highest_accuracy":14        k = int(len(corr) * 0.8)15    elif preference == "fastest":16        k = int(len(corr) * 0.3)17    else:18        k = int(len(corr) * 0.5)19 20    selected_cols = corr.head(max(k, 1)).index21    return df[selected_cols]