CUSTOMER CHURN PREDICTION ON E-COMMERCE PLATFORMS USING ENSEMBLE LEARNING AND BAYESIAN OPTIMIZATION METHODS
Keywords:
Customer Churn Prediction, Ensemble Learning, Bayesian Optimization, Machine Learning, E-CommerceAbstract
Customer churn is a critical issue in the e-commerce industry as it impacts customer loyalty and potential business losses. Therefore, a predictive model capable of accurately identifying customers with the potential to churn is needed. This study aims to analyze the performance of ensemble learning methods in predicting customer churn and evaluate the effect of Bayesian Optimization on improving model performance. The methods used include three ensemble algorithms, namely Random Forest, Gradient Boosting, and XGBoost, with a hyperparameter optimization process using Optuna-based Bayesian Optimization. The dataset is divided into 80% training data (3152 data) and 20% testing data (789 data) with a total of 10 main features. Model evaluation was carried out using accuracy, precision, recall, F1-score, and ROC-AUC metrics. The results showed that all models experienced improved performance after hyperparameter optimization. The XGBoost model provided the best results with an accuracy of 95.44%, precision of 91.60%, recall of 80.74%, F1-score of 85.83%, and ROC-AUC of 95.55%. Feature importance analysis showed that Tenure, Complain, and NumberOfAddress variables were the most influential factors in predicting customer churn. Thus, the combination of ensemble learning and Bayesian Optimization proved effective in improving the accuracy of customer churn prediction and can be used as a decision support for e-commerce companies in developing customer retention strategies.
References
[1] M. Iqbal, F. Nuzula, and C. Fatichah, “Optimasi hyperparameter pada convolutional neural network untuk klasifikasi jenis penyakit kanker kulit menggunakan Bayesian optimization,” vol. 7, no. 1, pp. 146–157, 2025.
[2] A. Iskoko and I. Tahyudin, “Hyperparameter optimization of IndoBERT using grid search, random search, and Bayesian optimization in sentiment analysis of e-government application reviews,” vol. 6, no. 5, pp. 3430–3444, 2025.
[3] H. Mulyo and A. K. Zyen, “Pengaruh hyperparameter tuning gradient boosting terhadap prediksi pemilihan program studi mahasiswa baru,” Bulletin of Computer Science Research, vol. 5, no. 2, pp. 131–137, 2025.
[4] A. Rama, P. Astawa, and G. H. Martono, “Penerapan ensemble learning dengan hard voting untuk klasifikasi customer churn,” Sep. 2025.
[5] F. R. Salwa, M. Novita, and R. Renaldy, “Peningkatan performa prediksi survival pasien gagal jantung menggunakan stacking ensemble learning,” vol. 3, pp. 614–625, 2025.
[6] I. Saviola, T. Agung, and B. Wirayuda, “Customer churn prediction pada streaming music platform menggunakan ensemble learning,” vol. 3, no. 2, pp. 76–82, 2025.
[7] N. Y. Arifin, “Studi metodologis optimasi hyperparameter XGBoost menggunakan Bayesian optimization untuk prediksi risiko stunting berbasis dataset simulasi,” vol. 3, no. 1, pp. 41–48, 2026.
[8] J. Informatika, F. Sains, U. Jenderal, and A. Yani, “Klasifikasi cuaca Jawa Barat menggunakan ensemble learning pada data meteorologi,” Weather Classification in West Java Using Ensemble Learning on Meteorological Data, vol. 14, pp. 2028–2044, 2025.
[9] S. Kania, N. Alya, I. Jumiatin, and I. Sulistia, “Implementation of hyperparameter tuning for classification models in heart disease risk prediction,” vol. 5, pp. 1181–1189, Oct. 2025.
[10] E. J. Kusuma, R. Nurmandhani, L. Aryani, I. Pantiawati, and G. F. Shidik, “Optimasi model extreme gradient boosting dalam upaya penentuan tingkat risiko pada ibu hamil,” vol. 12, no. 1, 2025, doi: 10.25126/jtiik.2025129001.
[11] I. Maulana and S. Ernawati, “Meningkatkan klasifikasi penyakit diabetes menggunakan metode ensemble soft voting dengan SMOTE-ENN dan optimasi Bayesian,” vol. 13, no. 1, pp. 71–86, 2025.
[12] M. Sakmar, E. D. Ramadhany, S. Hartinah, and R. A. Surya, “Hyperparameter optimisation in XGBoost for air quality prediction in Indonesian urban cities,” vol. 6, no. 1, pp. 48–56, 2025, doi: 10.24176/detika.v6i1.159625.
[13] M. P. Server, “Pengembangan model stacking machine learning dengan,” Jurnal Sains Informatika Terapan (JSIT), pp. 190–194, 2026.
[14] N. L. Wicaksono, “Optimasi hyperparameter ensemble learning untuk prediksi penyakit jantung berdasarkan data pasien,” Hyperparameter Optimization of Ensemble Learning for Heart Disease Prediction Using Patient Data, vol. 15, pp. 896–906, 2026.
[15] A. Yuandhika and M. Idhom, “Implementasi extremely randomized trees dengan optimasi hyperparameter accelerated particle swarm optimization untuk klasifikasi subtipe anemia,” vol. 16, no. 1, pp. 46–56, 2026.
