basemodel/train_randomforest.py

import pandas as pd
import joblib
from sklearn.ensemble import RandomForestClassifier
from sklearn.metrics import accuracy_score, classification_report


train_df = pd.read_csv("basemodel/train_features.csv")
val_df = pd.read_csv("basemodel/val_features.csv")


drop_cols = ["Label"]
if "language" in train_df.columns:
    drop_cols.append("language")

X_train = train_df.drop(columns=drop_cols)
y_train = train_df["Label"]

X_val = val_df.drop(columns=drop_cols)
y_val = val_df["Label"]


rf = RandomForestClassifier(
    n_estimators=200,
    max_depth=8,
    min_samples_split=5,
    min_samples_leaf=3,
    random_state=42,
    class_weight="balanced"
)

rf.fit(X_train, y_train)


val_preds = rf.predict(X_val)

accuracy = accuracy_score(y_val, val_preds)
print("\nValidation Accuracy:", round(accuracy, 4))

print("\nValidation Classification Report:\n")
print(classification_report(y_val, val_preds, target_names=["Human", "AI"]))


joblib.dump(rf, "basemodel/random_forest_baseline.pkl")
print("\n✅ Random Forest baseline model saved to basemodel/random_forest_baseline.pkl")