PREDICTING BOND PRICE MOVEMENTS USING A HYBRID METHOD BASED ON XGBOOST AND A GENETIC ALGORITHM
Abstract
The article presents a hybrid method for predicting the direction of bond price movements, combining the XGBoost machine learning method with hyperparameter optimization using a genetic algorithm. The research is aimed at solving the problem of binary classification of the direction of the price of the Russian Railways bond on the next trading day. The research methodology includes the formation of an expanded feature space of 18 technical indicators calculated on the basis of daily OHLCV data. To configure XGBoost hyperparameters, a genetic algorithm is implemented using the DEAP library. The study was conducted on three time horizons: 01.01.21-31.10.25, 01.01.23-31.10.25, from 01.01.24-31.10.25. As a result, a significant dependence of the effectiveness of the model on the time horizon of the training data is shown. The best quality was demonstrated by a model trained on data from 2024-2025, with an accuracy of 64.4% in the test sample, balanced precision and recall metrics, as well as high F1-score scores for both classes. Models trained over longer periods (2021-2025 and 2023-2025) showed a decrease in generalizing ability, which indicates that the relevance of the data prevails over its volume in the context of changes in Russia's monetary policy in 2021-2025. To maintain the predictive power of the model in changing market conditions, it is recommended to use a sliding learning window of 1.5–2 years. The comparison with the "Buy & Hold" strategy confirmed the effectiveness of the proposed hybrid approach
##article.references##
1. Gu Shihao & Kelly Bryan & Xiu Dacheng. Empirical Asset Pricing via Machine Learning, The Review of Financial Studies, 2020, 33, pp. 2223-2273. 10.1093/rfs/hhaa009.
2. Cabrol Axel & Drobetz Wolfgang & Otto Tizian & Puhan Tatjana. Predicting Corporate Bond Illiquidi-ty via Machine Learning, SSRN Electronic Journal, 2023. 10.2139/ssrn.4489504.
3. Mishin A.A., Vakulenko O.S. Integrated approach to forecasting government bond prices based on event analysis and machine learning, Economics and management: problems, solutions. Moscow, 2025, Vol. 3, No. 7, pp. 165-174.
4. Chen Tianqi & Guestrin Carlos. XGBoost: A Scalable Tree Boosting System, 2016, pp. 785-794. 10.1145/2939672.2939785.
5. Makarov D.A. Ispol'zovanie algoritma XGBoost dlya predskazaniya zaversheniya kursa obuchayush-chimsya [Using the XGBoost algorithm to predict course completion for students], StudNet, 2021, No. 1. Available at: https://cyberleninka.ru/article/n/ispolzovanie-algoritma-xgboost-dlya-predskazaniya-zaversheniya-kursa-obuchayuschimsya.
6. Syahputra Akhmal & Saputro Rujianto. Application of the XGBoost Model with Hyperparameter Tun-ing for Industry Classification for Job Applicants, 2024. SinkrOn 8, pp. 1920-1931. 10.33395/sinkron.v8i3.13840.
7. Yan Q., Hu H. XGBoost Machine Learning Model and Bond Pricing, SSRN, 2024. Available at: https://papers.ssrn.com/sol3/papers.cfm?abstract_id=4846755.
8. Salazov A.V. Sravnitel'nyy analiz effektivnosti XGboost i LSTM dlya prognozirovaniya volatil'nosti aktsiy Sberbanka [Comparative analysis of the effectiveness of XGBoost and LSTM for predicting the volatility of Sberbank shares], Auditorium, 2025, No. 2 (46). Available at: https://cyberleninka.ru/article/ n/sravnitelnyy-analiz-effektivnosti-xgboost-i-lstm-dlya-prognozirovaniya-volatilnosti-aktsiy-sberbanka.
9. Bakumenko L.P., Vasil'eva N.S. Sravnitel'nyy analiz prognosticheskikh modeley razvitiya kriptovalyut-nogo rynka (LSTM i XGBOOST) na primere bitkoina [Comparative analysis of predictive models of the development of the cryptocurrency market (LSTM and XGBOOST) on the example of bitcoin], Uchet i statistika [Accounting and Statistics], 2023, No. 4 (72). Available at: https://cyberleninka.ru/article/n/sravnitelnyy-analiz-prognosticheskih-modeley-razvitiya-kriptovalyutnogo-rynka-lstm-i-xgboost-na-primere-bitkoina.
10. Zhang Lingyu & Bian Wenjie & Qu Wenyi & Tuo Liheng & Wang Yunhai. Time series forecast of sales volume based on XGBoost, Journal of Physics: Conference Series, 2021, 1873, 012067. 10.1088/1742-6596/1873/1/012067.
11. Thejovathi Murari and Rao, Dr. M V P Chandra Sekhara and Priyadharsini E.J. and Siddi Someshwar and Karthik B. and Abbas Syed Hauider. Optimizing Product Demand Forecasting with Hybrid Ma-chine Learning and Time Series Models: A Comparative Analysis of XGBoost and SARIMA (Novem-ber 15, 2024), Proceedings of the 3rd International Conference on Optimization Techniques in the Field of Engineering (ICOFE-2024).
12. Nichani Rabia & Laid Gasmi & Laiche Nabil & Salheddine Kabou. Optimizing financial time series predictions with hybrid ARIMA, LSTM, and XGBoost Models, Studies in Engineering and Exact Sci-ences, 2024, 5. e11188. 10.54021/seesv5n2-582.
13. Odhiambo Stephen & Nyakundi Cornelious & Waititu Hellen. Developing a Hybrid ARIMA-XGBOOST Model for Analysing Mobile Money Transaction Data in Kenyam, Asian Journal of Prob-ability and Statistics, 2024, 26, pp. 108-126. 10.9734/ajpas/2024/v26i10662.
14. Gond B. PHM-MAX-XGBoost-LSTM-SVM: A Hybrid Model for Time Series Forecasting – A Com-parative Analysis with Traditional and Modern Approaches, 2025. 10.31224/4754.
15. Adil Mehdary, Abdellah Chehri, Abdeslam Jakimi, Rachid Saadane. Hyperparameter Optimization with Genetic Algorithms and XGBoost: A Step Forward in Smart Grid Fraud Detection, Sensors, 2024. Available at: https://www.mdpi.com/1424-8220/24/4/1230.
16. Lingeswari R., Brindha S. Online payments fraud prediction using optimized genetic algorithm based feature extraction and modified loss with XGboost algorithm for classification, Swarm and Evolutionary Computation, Vol. 95. Available at: https://doi.org/10.1016/j.swevo.2025.101934.
17. Kyung Keun Yun, Sang Won Yoon, Daehan Won. Prediction of stock price direction using a hybrid GA-XGBoost algorithm with a three-stage feature engineering process, Expert Systems with Applications, 2021, Vol. 186, 115716. ISSN 0957-4174. Available at: https://doi.org/10.1016/j.eswa.2021.115716.
18. Yun K.K., et al. Prediction of stock price direction using a hybrid GA–XGBoost algorithm with a three-stage feature engineering process, Expert Systems with Applications, 2021. Available at: https://www.sciencedirect.com/science/article/abs/pii/S0957417421010988.
19. A Guide on XGBoost hyperparameters tuning. Available at: https://www.kaggle.com/code/ prashant111/a-guide-on-xgboost-hyperparameters-tuning.
20. DEAP documentation. Available at: https://deap.readthedocs.io/en/master/.
21. Shvager D. Tekhnicheskiy analiz: polnyy kurs: perevod s angliyskogo [Technical analysis: a complete course]. Moscow: Al'pina Pablisher, 2021, 804 p.
