Abstract

This study proposes a hybrid forecasting model that integrates Neural Networks and Linear Regression to address complex financial prediction tasks involving gold prices, Bitcoin values, and USD/TRY exchange rates. The architecture includes a linear regression component for modeling direct relationships using flattened input vectors, and an LSTM component with dense layers to capture nonlinear temporal dependencies from three-dimensional sequential data. SHAP (SHapley Additive exPlanations) is employed to enhance model explainability by identifying the most influential features. Additionally, an AI-powered chatbot assistant enables users to access forecasts and receive natural language explanations interactively. Experimental results show that the hybrid model significantly reduces prediction error, achieving a mean absolute percentage error of 5.5%. The integration of explainability tools and user interaction enhances model transparency and usability. Overall, the proposed approach offers a robust and interpretable framework for financial time series forecasting.

References

1. Livieris, I. E., Pintelas, E., & Pintelas, P. A CNN–LSTM model for gold price time-series forecasting. Neural Computing and Applications, (2020). 32(15), 17351–17360.

2. Sadorsky, P. Predicting Gold and Silver Price Direction Using Tree-Based Classifiers. Journal of Risk and Financial Management, (2021). 14(5), 198.

3. Cao, W., Zhu, W., Wang, W., Demazeau, Y., & Zhang, C. A deep coupled LSTM approach for USD/CNY exchange rate forecasting. IEEE Intelligent Systems, (2020). 35(2), 43–53.

4. Karasu, S., Altan, A., Saraç, Z., & Hacıoğlu, R. Prediction of Bitcoin prices with machine learning methods using time series data. In 2018 26th Signal Processing and Communications Applications Conference (SIU) (2018). (pp. 1–4). IEEE.

5. Li, Y., & Dai, W. Bitcoin price forecasting method based on CNN-LSTM hybrid neural network model. Journal of Engineering, 2020(1), 344–347.

6. Zhang, G.P. Time series forecasting using a hybrid ARIMA and neural network model. Neurocomputing, (2003) 50(C), 159–175.

7. Das, S. R., Mishra, D., & Rout, M. A hybridized ELM using self-adaptive multi-population-based Jaya algorithm for currency exchange prediction: an empirical assessment. Neural Computing and Applications, (2019). 31(11), 7071–7094.

8. Şimşek, A. İ. Performance comparison of machine and deep learning methods in USD/TRY exchange rate forecasting. Akademik Yaklaşımlar Dergisi (Journal of Academic Approaches), (2024). 15(3), 1473–1499.

9. Clavería, O., Monte, E., Sorić, P., & Porras, S. An application of deep learning for exchange rate forecasting. SSRN Electronic Journal. (2022).

10. Lundberg, S. M., & Lee, S. I. A Unified Approach to Interpreting Model Predictions. In Advances in Neural Information Processing Systems (2017). 30 (NIPS 2017).

11. Ribeiro, M. T., Singh, S., & Guestrin, C. “Why Should I Trust You?”: Explaining the Predictions of Any Classifier. In Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining (2016). (pp. 1135–1144)

12. Choi, J. E., Shin, J. W., & Shin, D. W. Vector SHAP Values for Machine Learning Time Series Forecasting. Journal of Forecasting (2025). 44(2), 635–645.

13. Sen, D., Deora, B. S., & Vaishnav, A. Explainable Deep Learning for Time Series Analysis: Integrating SHAP and LIME in LSTM-Based Models. Journal of Information Systems Engineering and Management (2025)., 10(SI), Article 16.

14. Sun, S., Wang, S., & Wei, Y. A new ensemble deep learning approach for exchange rates forecasting and trading. Advanced Engineering Informatics, (2020). 46, 101160.

15. Brown, T., et al. Language Models are Few-Shot Learners. Advances in Neural Information Processing Systems (2020). (NeurIPS 2020).

16. See, A., Liu, P. J., & Manning, C. D. Get To The Point: Summarization with Pointer-Generator Networks. Proceedings of ACL 2019 (2019).