Perbandingan XGBoost dan LSTM untuk Prediksi Rentang Harga Penutupan Saham BBRI, ANTM dan ASII

Auliya Afifah Adnan Hakim; Adha Mashur Sajiah; Ilham Julian Efendi

doi:10.69693/ijim.v4i3.2030

Authors

Auliya Afifah Adnan Hakim Universitas Halu Oleo
Adha Mashur Sajiah Universitas Halu Oleo
Ilham Julian Efendi Universitas Halu Oleo

DOI:

https://doi.org/10.69693/ijim.v4i3.2030

Keywords:

XGBoost, LSTM, Harga Saham, Regresi Kuantil, Kondisi Pasar

Abstract

Prediksi harga saham harian umumnya menghasilkan satu nilai titik sehingga tidak menggambarkan ketidakpastian pergerakan harga, padahal pelaku pasar membutuhkan gambaran rentang untuk mengelola risiko. Penelitian ini bertujuan membandingkan kinerja XGBoost dan Long Short-Term Memory (LSTM) dalam memprediksi titik tengah serta rentang harga penutupan harian saham BBRI, ASII, dan ANTM. Data historis 2020 sampai 2025 dibentuk menjadi empat belas fitur relatif, lalu dibagi secara kronologis menjadi data training, validation, dan testing agar informasi masa depan tidak bocor ke dalam proses pelatihan. Ketiga kuantil, yaitu kuantil bawah, tengah, dan atas, dipelajari secara terpisah melalui pendekatan regresi kuantil. Evaluasi dilakukan dengan akar galat kuadrat rata-rata, galat persentase absolut rata-rata, koefisien determinasi, ketepatan arah, cakupan rentang, lebar rentang rata-rata, interval score, waktu komputasi, dan ukuran model. Ketahanan model juga diuji pada kondisi bullish, bearish, dan sideways, kemudian diverifikasi melalui forward test 121 hari perdagangan tahun 2026 tanpa pelatihan ulang. Pada historical test, XGBoost memberikan RMSE terendah untuk seluruh saham dan unggul pada tujuh dari sembilan kombinasi kondisi pasar. Pada forward test, XGBoost tetap unggul untuk BBRI dan ASII, sedangkan LSTM sedikit lebih baik untuk ANTM serta menghasilkan coverage lebih tinggi pada ketiga saham. Hasil ini menunjukkan bahwa XGBoost lebih konsisten untuk prediksi titik, sementara LSTM lebih baik dalam mencakup harga aktual pada rentang prediksi forward.

Downloads

Download data is not yet available.

References

Aiyegbeni, G. & Li, Y. 2024. A Comparative Analysis of LSTM, ARIMA, XGBoost Algorithms in Predicting Stock Price Direction. Engineering and Technology Journal. Vol. 9, No. 8, hlm. 4978-4986. https://doi.org/10.47191/etj/v9i08.50

Akiba, T., Sano, S., Yanase, T., Ohta, T. & Koyama, M. 2019. Optuna: A Next-Generation Hyperparameter Optimization Framework. Proceedings of the 25th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining. hlm. 2623-2631. https://doi.org/10.1145/3292500.3330701

Chen, T. & Guestrin, C. 2016. XGBoost: A Scalable Tree Boosting System. Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining. hlm. 785-794. https://doi.org/10.1145/2939672.2939785

Chicco, D., Warrens, M. J. & Jurman, G. 2021. The Coefficient of Determination R-Squared is More Informative than SMAPE, MAE, MAPE, MSE and RMSE in Regression Analysis Evaluation. PeerJ Computer Science. Vol. 7, art. e623. https://doi.org/10.7717/peerj-cs.623

Gneiting, T. & Raftery, A. E. 2007. Strictly Proper Scoring Rules, Prediction, and Estimation. Journal of the American Statistical Association. Vol. 102, No. 477, hlm. 359-378. https://doi.org/10.1198/016214506000001437

Hochreiter, S. & Schmidhuber, J. 1997. Long Short-Term Memory. Neural Computation. Vol. 9, No. 8, hlm. 1735-1780. https://doi.org/10.1162/neco.1997.9.8.1735

Karunasingha, D. S. K. 2022. Root Mean Square Error or Mean Absolute Error? Use Their Ratio as Well. Information Sciences. Vol. 585, hlm. 609-629. https://doi.org/10.1016/j.ins.2021.11.036

Koenker, R. & Bassett, G. 1978. Regression Quantiles. Econometrica. Vol. 46, No. 1, hlm. 33-50. https://doi.org/10.2307/1913643

Kumbure, M. M., Lohrmann, C., Luukka, P. & Porras, J. 2022. Machine Learning Techniques and Data for Stock Market Forecasting: A Literature Review. Expert Systems with Applications. Vol. 197, art. 116659. https://doi.org/10.1016/j.eswa.2022.116659

Li, Z. 2023. Comparison of XGBoost and LSTM Models for Stock Price Prediction. Advances in Economics, Management and Political Sciences. Vol. 61, hlm. 147-155. https://doi.org/10.54254/2754-1169/61/20231181

Lin, Y. H., Liu, S. C., Yang, H. J. & Wu, H. R. 2021. Stock Trend Prediction Using Candlestick Charting and Ensemble Machine Learning Techniques with a Novelty Feature Engineering Scheme. IEEE Access. Vol. 9, hlm. 101433-101446. https://doi.org/10.1109/ACCESS.2021.3096825

Liu, F., Umair, M. & Gao, J. 2023. Assessing Oil Price Volatility Co-Movement with Stock Market Volatility Through Quantile Regression Approach. Resources Policy. Vol. 81, art. 103375. https://doi.org/10.1016/j.resourpol.2023.103375

Makridakis, S., Spiliotis, E. & Assimakopoulos, V. 2022. M5 Accuracy Competition: Results, Findings, and Conclusions. International Journal of Forecasting. Vol. 38, No. 4, hlm. 1346-1364. https://doi.org/10.1016/j.ijforecast.2021.11.013

Mostafavi, S. M. & Hooman, A. R. 2025. Key Technical Indicators for Stock Market Prediction. Machine Learning with Applications. Vol. 20, art. 100631. https://doi.org/10.1016/j.mlwa.2025.100631

Oksak, Y., Buyukkor, Y. & Saritas, T. 2025. Wavelet-Enhanced Multimodel Framework for Stock Market Forecasting: A Comprehensive Analysis Across Market Regimes. Borsa Istanbul Review. art. 100771. https://doi.org/10.1016/j.bir.2025.100771

Oukhouya, H., Kadiri, H., El Himdi, K. & Guerbaz, R. 2024. Forecasting International Stock Market Trends: XGBoost, LSTM, LSTM-XGBoost, and Backtesting XGBoost Models. Statistics, Optimization and Information Computing. Vol. 12, No. 1, hlm. 200-209. https://doi.org/10.19139/soic-2310-5070-1822

Pratama, B. & Banowosari, L. Y. 2024. Perbandingan Metode Extreme Gradient Boosting (XGBoost) dengan Long Short-Term Memory (LSTM) untuk Prediksi Saham PT Bank Mandiri Tbk. (BMRI). Journal of Economic, Business and Accounting. Vol. 7, No. 3, hlm. 5631-5636. https://doi.org/10.31539/costing.v7i3.9473

Raudys, A. & Goldstein, E. 2022. Forecasting Detrended Volatility Risk and Financial Price Series Using LSTM Neural Networks and XGBoost Regressor. Journal of Risk and Financial Management. Vol. 15, No. 12, art. 602. https://doi.org/10.3390/jrfm15120602

Sathiyapriya, K., Vankadara, S., Babu, K. S. & Muralidharan, M. 2023. Performance Comparison of LSTM and XGBoost for Ether Price Prediction from Spam-Filtered Tweets. Proceedings of the International Conference on Intelligent Systems for Communication, IoT and Security. hlm. 650-655. https://doi.org/10.1109/ICISCoIS56541.2023.10100425

Selayanti, N., Putri, D. A., Trimono & Idhom, M. 2025. Prediksi Harga Penutupan Saham BBRI dengan Model Hybrid LSTM-XGBoost. Informatika: Jurnal Teknik Informatika dan Multimedia. Vol. 5, No. 1, hlm. 52-64. https://doi.org/10.51903/informatika.v5i1.1011

Xu, J., He, J., Gu, J., Wu, H., Wang, L., Zhu, Y., Wang, T., He, X. & Zhou, Z. 2022. Financial Time Series Prediction Based on XGBoost and Generative Adversarial Networks. International Journal of Circuits, Systems and Signal Processing. Vol. 16, hlm. 637-645. https://doi.org/10.46300/9106.2022.16.79

Yun, K. K., Yoon, S. W. & Won, D. 2021. Prediction of Stock Price Direction Using a Hybrid GA-XGBoost Algorithm with a Three-Stage Feature Engineering Process. Expert Systems with Applications. Vol. 186, art. 115716. https://doi.org/10.1016/j.eswa.2021.115716

Zeng, X., Cai, J., Liang, C. & Yuan, C. 2023. Prediction of Stock Price Movement Using an Improved NSGA-II-RF Algorithm with a Three-Stage Feature Engineering Process. PLOS ONE. Vol. 18, No. 6, art. e0287754. https://doi.org/10.1371/journal.pone.0287754