Evaluating a Geopolitical-Regime-Aware Hybrid Ensemble for IHSG Forecasting

Imam Prasodjo

doi:10.25134/ilkom.v20i2.654

Authors

Imam Prasodjo Universitas Al-Khairiyah Cilegon

DOI:

https://doi.org/10.25134/ilkom.v20i2.654

Keywords:

Geopolitical risk, hybrid ensemble, IHSG, LSTM, Machine learning, stock-price forecasting, XGBoost

Abstract

This study evaluates a geopolitical-regime-aware hybrid ensemble (GRAHE) for one-day-ahead forecasting of the Indonesia Composite Index (IHSG) using observations from 2015 to 2025. The framework combines long short-term memory, extreme gradient boosting, and random forest forecasts through a softmax gate conditioned on one-month-lagged geopolitical risk and rolling market volatility. Hyperparameters were selected with Bayesian optimization on the 2022 validation period, and the locked test used monthly expanding-window retraining from 2 January 2023 to 5 December 2025 (697 forecasts). GRAHE obtained a return RMSE of 0.009334, directional accuracy of 50.50%, and Theil’s U of 1.001. It did not outperform the strongest benchmark: Ridge regression achieved an RMSE of 0.009316, while XGBoost produced the highest directional accuracy of 54.23%. The RMSE difference between Ridge and GRAHE was not statistically significant according to the Diebold–Mariano test (p = 0.744). In high-GPR observations, GRAHE recorded an RMSE of 0.009626, 0.30% above the best baseline. Removing GPR features increased ensemble RMSE slightly to 0.009350, but market-only XGBoost remained competitive at 0.009326. These findings do not support a strong claim that geopolitical conditioning improves daily IHSG point forecasts, although the framework provides a transparent basis for regime-dependent risk analysis and further robustness testing.

Downloads

Download data is not yet available.

References

[1] D. Caldara and M. Iacoviello, “Measuring geopolitical risk,” American Economic Review, vol. 112, no. 4, pp. 1194–1225, 2022, doi: 10.1257/aer.20191823.

[2] A. A. Salisu, A. Ogbonna, L. Lasisi, and A. Olaniran, “Geopolitical risk and stock market volatility in emerging markets: A GARCH–MIDAS approach,” The North American Journal of Economics and Finance, vol. 62, art. 101755, 2022, doi: 10.1016/j.najef.2022.101755.

[3] D. Das, M. Kannadhasan, and M. Bhattacharyya, “Do the emerging stock markets react to international economic policy uncertainty, geopolitical risk and financial stress alike?” The North American Journal of Economics and Finance, vol. 48, pp. 1–19, 2019, doi: 10.1016/j.najef.2019.01.008.

[4] M. Yang, J. Zhang, and H. Li, “Geopolitical risk and stock market volatility in emerging economies: Evidence from GARCH-MIDAS model,” Discrete Dynamics in Nature and Society, art. 1159358, 2021, doi: 10.1155/2021/1159358.

[5] T. Chen and C. Guestrin, “XGBoost: A scalable tree boosting system,” in Proc. 22nd ACM SIGKDD Int. Conf. Knowledge Discovery and Data Mining, 2016, pp. 785–794, doi: 10.1145/2939672.2939785.

[6] S. Hochreiter and J. Schmidhuber, “Long short-term memory,” Neural Computation, vol. 9, no. 8, pp. 1735–1780, 1997, doi: 10.1162/neco.1997.9.8.1735.

[7] L. Breiman, “Random forests,” Machine Learning, vol. 45, pp. 5–32, 2001, doi: 10.1023/A:1010933404324.

[8] M. Nabipour, P. Nayyeri, H. Jabani, S. S., and A. Mosavi, “Deep learning for stock market prediction,” Entropy, vol. 22, no. 8, art. 840, 2020, doi: 10.3390/e22080840.

[9] P. H. Vuong, T. T. Dat, T. K. Mai, and P. H. Uyen, “Stock-price forecasting based on XGBoost and LSTM,” Computer Systems Science and Engineering, vol. 40, no. 1, pp. 237–246, 2022, doi: 10.32604/CSSE.2022.017685.

[10] Z. Shi, Y. Hu, G. Mo, and J. Wu, “Attention-based CNN-LSTM and XGBoost hybrid model for stock prediction,” arXiv:2204.02623, 2022, doi: 10.48550/arXiv.2204.02623.

[11] K. K. Yun, S. W. Yoon, and D. Won, “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, 2021, doi: 10.1016/j.eswa.2021.115716.

[12] H. Oukhouya, H. Kadiri, K. El Himdi, and R. Guerbaz, “Forecasting international stock market trends: XGBoost, LSTM, LSTM-XGBoost, and backtesting XGBoost models,” Statistics, Optimization & Information Computing, vol. 12, no. 1, pp. 200–209, 2024, doi: 10.19139/soic-2310-5070-1822.

[13] T. Akiba, S. Sano, T. Yanase, T. Ohta, and M. Koyama, “Optuna: A next-generation hyperparameter optimization framework,” in Proc. 25th ACM SIGKDD Int. Conf. Knowledge Discovery and Data Mining, 2019, pp. 2623–2631, doi: 10.1145/3292500.3330701.

[14] F. X. Diebold and R. S. Mariano, “Comparing predictive accuracy,” Journal of Business & Economic Statistics, vol. 13, no. 3, pp. 253–263, 1995, doi: 10.1080/07350015.1995.10524599.

[15] M. H. Pesaran and A. Timmermann, “A simple nonparametric test of predictive performance,” Journal of Business & Economic Statistics, vol. 10, no. 4, pp. 461–465, 1992, doi: 10.1080/07350015.1992.10509922.

[16] S. M. Lundberg and S.-I. Lee, “A unified approach to interpreting model predictions,” in Advances in Neural Information Processing Systems 30, 2017, pp. 4765–4774.

[17] R. J. Hyndman and A. B. Koehler, “Another look at measures of forecast accuracy,” International Journal of Forecasting, vol. 22, no. 4, pp. 679–688, 2006, doi: 10.1016/j.ijforecast.2006.03.001.

[18] R. F. Engle, “Autoregressive conditional heteroscedasticity with estimates of the variance of United Kingdom inflation,” Econometrica, vol. 50, no. 4, pp. 987–1007, 1982, doi: 10.2307/1912773.

[19] T. Bollerslev, “Generalized autoregressive conditional heteroskedasticity,” Journal of Econometrics, vol. 31, no. 3, pp. 307–327, 1986, doi: 10.1016/0304-4076(86)90063-1.

[20] S. Deng, X. Huang, Y. Zhu, Z. Su, Z. Fu, and T. Shimada, “Stock index direction forecasting using an explainable eXtreme Gradient Boosting and investor sentiments,” The North American Journal of Economics and Finance, vol. 64, art. 101848, 2023, doi: 10.1016/j.najef.2022.101848.

[21] S. Mehtab, J. Sen, and A. Dutta, “Stock price prediction using machine learning and LSTM-based deep learning models,” in Proc. Int. Conf. Decision Aid Sciences and Application, 2020, pp. 447–453.

[22] Z. Shi, Y. Hu, G. Mo, and J. Wu, “Source code for attention-based CNN-LSTM and XGBoost stock prediction,” associated repository, accessed for methodological comparison.

[23] M. Iacoviello, “Geopolitical Risk (GPR) Index: Data and replication materials,” official project data page, accessed July 2026.

[24] Indonesia Stock Exchange, “Composite Stock Price Index and market statistics,” official market-data portal, accessed July 2026.

[25] A. A. Salisu et al., “Geopolitical risk and stock market volatility: A global perspective,” Finance Research Letters, vol. 53, art. 103620, 2023, doi: 10.1016/j.frl.2022.103620.

[26] G. Harrison, “Daily IHSG: Daily Jakarta Composite Index historical data derived from Yahoo Finance,” Hugging Face Datasets, file ihsg_daily.csv, accessed July 30, 2026.