MODEL REKOMENDASI LAGU BERBASIS GENRE MENGGUNAKAN METODE RANDOM FOREST DAN DECISION TREE 

GENRE-BASED SONG RECOMMENDATION MODEL USING RANDOM FOREST AND DECISION TREE METHODS 

Penelitian ini mengembangkan model sistem rekomendasi lagu berbasis genre menggunakan algoritma Random Forest dan Decision Tree. Proses pemodelan dimulai dengan analisis feature importance untuk mengidentifikasi sepuluh fitur audio utama yang paling berpengaruh terhadap klasifikasi genre. Evaluasi performa dilakukan menggunakan teknik cross-validation guna memastikan hasil yang konsisten dan dapat digeneralisasi. Hasil pengujian menunjukkan bahwa algoritma Random Forest memberikan performa yang lebih unggul dibandingkan Decision Tree, ditunjukkan oleh nilai akurasi, presisi, recall, dan F1-score yang lebih tinggi. Secara kuantitatif, model Random Forest mencatat rata-rata akurasi sebesar 83%, sedangkan Decision Tree hanya mencapai 79%. Keunggulan ini menegaskan bahwa Random Forest merupakan metode yang lebih efektif dan andal untuk digunakan dalam membangun model rekomendasi lagu berdasarkan genre.

Kata Kunci : Sistem Rekomendasi, Genre Musik, Random Forest, Decision Tree, Feature Importance, Cross-Validation, Machine Learning. 

This study developed a genre-based song recommendation system model using the Random Forest and Decision Tree algorithms. The modeling process began with feature importance analysis to identify the ten key audio features most influential in genre classification. Performance evaluation was conducted using cross-validation techniques to ensure consistent and generalizable results. Test results showed that the Random Forest algorithm provided superior performance compared to Decision Tree, as indicated by higher accuracy, precision, recall, and F1 score values. Quantitatively, the Random Forest model recorded an average accuracy of 83%, while Decision Tree only achieved 79%. This superiority confirms that Random Forest is a more effective and reliable method for building genre-based song recommendation models.

Keywords : Recommender System, Music Genre, Random Forest, Decision Tree, Feature Importance, Cross-Validation, Machine Learning.