A Perspective on the Banking Sector with Deep Learning Method: Prediction of BIST Banking Index Movements

Year 2024, Volume: 13 Issue: 3, 1277 - 1291, 30.09.2024

Nazif Ayyıldız

https://doi.org/10.15869/itobiad.1451709

Abstract

Banks, the fundamental players in the financial system, play a crucial role in ensuring the healthy functioning of the economy. Bank indices are generally considered indicators of economic health, reflecting the performance of a country's financial sector. The BIST Bank Index, comprising leading bank stocks in Türkiye, represents the performance of the banking sector. On the other hand, predicting stock prices is often a complex issue influenced by various and variable factors. In addition to traditional methods such as fundamental and technical analysis used for forecasting in financial markets, numerous machine learning methods have been developed in recent years. Machine learning methods can effectively handle the non-linear and non-stationary characteristics of financial series, providing accurate predictions. Particularly, the deep learning method has gained prominence in prediction applications by efficiently processing large datasets and identifying non-linear relationships with high accuracy. The aim of this study is to predict the directional movements of the BIST Bank Index, which includes the leading bank stocks in Türkiye, using the deep learning method. The analysis incorporates weekly closing values of the BIST Bank Index from January 1, 2013, to December 31, 2023, along with weekly data on deposit and loan interest rates, overnight interest rates, deposit and loan volumes, total assets of the banking sector, exchange rates (USD and Euro), and BIST 100 index closing values. A total of 574 weeks of data were obtained for each input variable, resulting in the utilization of 5,740 financial data points in the analysis. The analysis revealed that the directional movements of the BIST Bank Index were predicted with an accuracy of 88.70% using the deep learning method. These findings demonstrate that the deep learning method can be effectively employed to predict the directional movements of bank indices with a certain level of accuracy.

Keywords

Financial Markets, Banks, BIST Bank Index, Deep Learning, Classification

Bankacılık Sektörüne Derin Öğrenme Yöntemiyle Bakış: BİST Banka Endeksi Hareket Yönlerinin Tahmini

Abstract

Finansal sistemdeki temel oyuncular olan bankalar, ekonominin sağlıklı işlemesinde kritik bir rol oynamaktadırlar. Banka endeksleri ise, genellikle bir ülkenin finansal sektöründeki performansı yansıtarak ekonomik sağlığın bir göstergesi olarak kabul edilmektedir. BIST Banka Endeksi, Türkiye'nin önde gelen banka hisselerini içeren bir endeks olup, bankacılık sektörünün performansını temsil etmektedir. Diğer yandan, hisse senedi fiyatlarının tahmin edilebilirliği, genellikle karmaşık ve değişken faktörlerle etkilenen bir konudur. Finansal piyasalarda tahmin amacıyla kullanılan temel analiz ve teknik analiz gibi geleneksel yöntemlere ek olarak, son dönemde çok sayıda makine öğrenimi yöntemi geliştirilmiştir. Makine öğrenimi yöntemleri, finansal serilerin doğrusal ve durağan olmayan özelliklerini ele alarak doğru tahminler yapabilmektedir. Tahmin uygulamalarındaki başarısı ile ön plana çıkan derin öğrenme yöntemi ise, büyük veri setlerini etkili bir şekilde işleyerek doğrusal olmayan ilişkileri belirlemekte ve yüksek doğrulukla çıkarım yapabilmektedir. Bu çalışmanın amacı, Türkiye'nin önde gelen banka hisselerini içeren BIST Banka Endeksi’nin hareket yönlerinin derin öğrenme yöntemi ile tahmin edilmesidir. Analizde, BIST Banka Endeksi'nin 01.01.2013-31.12.2023 dönemindeki haftalık kapanış değerleriyle birlikte, yine haftalık bazda elde edilen mevduat ve kredi faiz oranları, gecelik faiz oranları, mevduat ve kredi hacimleri, bankacılık sektörü aktif toplamı, döviz kurları (Dolar ve Euro) ve BIST 100 endeksi kapanış değerleri girdi verisi olarak kullanılmıştır. Her bir girdi değişkeni için 574 haftalık veri elde edilmiş olup toplam 5.740 adet veri analizde kullanılmıştır. Gerçekleştirilen analiz sonucunda, derin öğrenme yöntemi ile BIST Banka Endeksi’nin hareket yönleri %88,70 doğrulukta tahmin edilmiştir. Elde edilen bulgular, derin öğrenme yöntemi kullanılarak banka endeks hareket yönlerinin belirli bir seviyede doğru tahmin edilebileceğini göstermektedir.

Keywords

Finansal Piyasalar, Bankalar, BIST Banka Endeksi, Derin Öğrenme, Sınıflandırma

References

• Arora, A., Candel, A., Lanford, J., Ledel, E., & Parmar, V. (2015). Deep Learning with H2O. H2O.ai, Erişim Tarihi:22.02.2024, DeepLearning_Vignette.pdf(h2o-release.s3.amazonaws.com)
• Ayyıldız, N. (2023). Prediction of Stock Market Index Movements with Machine Learning. Özgür Publications. DOI: https://doi.org/10.58830/ozgur.pub354
• Beniwal, M., Singh, A. & Kumar, N. (2024). Forecasting Multistep Daily Stock Prices for Long-Term İnvestment Decisions: A Study of Deep Learning Models on Global Indices”, Engineering Applications of Artificial Intelligence, 129, DOI:10.1016/j.engappai.2023.107617
• Borsa İstanbul-BIST (2024), BIST Banka Endeks Bileşenleri, Erişim Tarihi: 18.02.2024, https://www.borsaistanbul.com/tr/endeks-detay/264/bist-banka
• Candel, A. & Ledel, E. (2015). Deep Learning with H2O. H2O.ai, Erişim Tarihi:27.02.2024, DeepLearning_Vignette.pdf (h2o-release.s3.amazonaws.com)
• Chaurasia, A. & Tiwari, R.K. (2021). Stock Price Prediction using Various Machine Learning. International Journal of Advances in Engineering and Management. 3(1):573-581, DOI: 10.35629/5252-0301573581
• Cui, C., Wang, P., Li, Y. & Zhang, Y. (2023), McVCsB: A New Hybrid Deep Learning Network for Stock İndex Prediction, Expert Systems with Applications, 232, DOI:10.1016/j.eswa.2023.120902
• Fama, E.F. (1970). Efficient Capital Markets: A Review of Theory and Empirical Work. Journal of Finance, 25 (2), 383-417, DOI:10.2307/2325486
• Gong, M. (2021). A Novel Performance Measure for Machine Learning Classification. International, Journal of Managing Information Technology, 13(1), 11–19. DOI:10.5121/IJMIT.2021.13101
• Gündüz, H., Cataltepe, Z. & Yaslan, Y. (2017). Stock Market Direction Prediction using Deep Neural Networks, 25th Signal Processing and Communications Applications Conference (SIU), Antalya, Türkiye, 1-4, DOI:10.1109/SIU.2017.7960512
• Gündüz, H., Yaslan, Y. & Çataltepe, Z (2018). Stock Market Prediction with Deep Learning Using Financial News, 26th IEEE Signal Processing and Communications Applications Conference (SIU), IEEE, New York, Amerika Birleşik Devletleri
• Gündüz, H (2020). Stock Market Prediction with Stacked Autoencoder Based Feature Reduction. 28th Signal Processing and Communications Applications Conference (SIU). DOI:10.1109/siu49456.2020.9302391
• H20 (2024), H20 Deep Learning (Neural Networks), Erişim Tarihi:25.02.2024, https://docs.h2o.ai/h2o/latest- stable/h2o-docs/data-science/deep-learning.html
• Jin, S. (2023). Sentiment-Driven Forecasting LSTM Neural Networks for Stock Prediction-Case of China Bank Sector, International Journal of Advanced Computer Science and Applications (IJACSA), 14(11), 2023. DOI:10.14569/IJACSA.2023.0141101
• Lin,S.W., Shiue,Y.R, Chen,S.C. & Cheng,H.M. (2009), Applying Enhanced Data Mining Approaches In Predicting Bank Performance: A Case of Taiwanese Commercial Banks, Expert Systems with Applications, 36(9), 11543- 11551, DOI:10.1016/j.eswa.2009.03.029.
• Kanwal, A., Lau,M.F., P.H. Ng,S., Sım,K.Y. & Chandrasekaran,S. (2022). BiCuDNNLSTM-1dCNN-A Hybrid Deep Learning-Based Predictive Model for Stock Price Prediction, Expert Systems with Applications, 202, DOI:10.1016/j.eswa.2022.117123.
• Kilimci Z. H. and Duvar, R. (2020). An Efficient Word Embedding and Deep Learning Based Model to Forecast the Direction of Stock Exchange Market Using Twitter and Financial News Sites: A Case of Istanbul Stock Exchange (BIST 100), IEEE Access, 8, 188186-188198, DOI:10.1109/ACCESS.2020.3029860
• 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, 197, DOI:10.1016/j.eswa.2022.116659.
• Liu, Y, Zhou, Y. Wen, S. & Tang, C. (2014), A Strategy on Selecting Performance Metrics for Classifier Evaluation”, International Journal of Mobile Computing and Multimedia Communication,, 6, 20-35 DOI:10.4018/IJMCMC.2014100102
• Mahajan, S. Anand, P. K. Sarangi & A. K. Sahoo, (2023) Escalate the Returns With ML-Based Technical Analysis: Next Day Closing Price Prediction Using RNN and LSTM Models, 3rd International Conference on Advance Computing and Innovative Technologies in Engineering (ICACITE), 261-266, DOI:10.1109/ICACITE57410.
• Mallıkarjuna, M.& Rao, R.P. (2019), Evaluation of Forecasting Methods from Selected Stock Market Returns. Financ Innov, 5, 40. DOI:10.1186/s40854-019-0157-x
• Mcwera, A. & Mba, J. C., (2023). Predicting Stock Market Direction in South African Banking Sector Using Ensemble Machine Learning Techniques. Data Science in Finance and Economics, 3(4): 401-426. DOI:10.3934/DSFE.2023023
• Najafabadı, M. M., Vıllanustre, F., Khoshgoftaar, T. M., Selıya, N., Wald, R., & Muharemagıc, E. (2015). Deep Learning Applications and Challenges in Big Data Analytics. Journal Of Big Data, 2(1), 1-21. DOI:10.1186/s40537- 014-0007-7
• Sokolova, M. & Lapalme, G. (2009). A Systematic Analysis of Performance Measures For Classification Tasks. Information Processing & Management. 45. 427-437. DOI:10.1016/j.ipm.2009.03.002
• Stetsenko, P. (2017), Machine Learning with Python and H2O, Fifth Edition, Erişim Tarihi: 28.02.2024, DeepLearning_Vignette.pdf (h2o-release.s3.amazonaws.com)
• Türkiye Cumhuriyet Merkez Bankası. (2024). TCMB Bankacılık Verileri, Erişim Tarihi:20.02.2024,https://www.tcmb.gov.tr/wps/wcm/connect/TR/TCMB+TR/Main+Menu/Istatistikler/Bankacil ik+Verileri/
• Zhong, X., & Enke, D. (2019). Predicting the Daily Return Direction Of The Stock Market Using Hybrid Machine Learning Algorithms, Financial Innovation, 5. DOI:10.1186/s40854-019-0138-0.