Likiditenin Derin Öğrenme Hatalarına Etkisi

Abstract

Bu çalışma, piyasa likiditesinin ve makroekonomik değişkenlerin finansal piyasalardaki derin öğrenme modellerinin tahmin performansı üzerindeki etkisini araştırmaktadır. Temel amaç, tek katmanlı bir Uzun Kısa Süreli Bellek (LSTM) modeli kullanarak BIST 30 endeksinde kote edilen on hisse senedine ilişkin fiyat hareketlerini tahmin etmek ve tahmin başarısızlıklarıyla ilişkili piyasa koşullarını tespit etmektir. 2006–2025 dönemini kapsayan günlük veriler kullanılarak iki aşamalı ampirik bir çerçeve uygulanmaktadır. İlk olarak, LSTM hiperparametreleri Bayesçi optimizasyon ve ileriye dönük doğrulama (walk-forward validation) yoluyla optimize edilmektedir. İkinci olarak, hisse senedi düzeyindeki likidite ölçütlerinin ve küresel makroekonomik değişkenlerin etkilerini incelemek üzere günlük Mutlak Yüzde Hata (APE), OLS-HAC ve EGARCH-X regresyonları kullanılarak modellenmektedir. Sonuçlar, LSTM modelinin on hisse senedinin tamamında %3.29 ile %8.40 arasında değişen MAPE değerleriyle iyi kalibre edilmiş tahminler ürettiğini göstermektedir. İşlem hacmi ve Amihud oranıyla ölçülen likidite düşüklüğü (illikidite), hisse senetlerinin çoğu için tahmin hatalarıyla pozitif yönde ilişkilidir. On hisse senedinden dokuzunda en az bir likidite göstergesi istatistiksel olarak anlamlıdır ve anlamlılığın gözlemlendiği durumlarda katsayı işaretleri tutarlılık sergilemektedir. Buna karşılık, makroekonomik ve küresel göstergeler sınırlı ve hisse senedine özgü etkiler ortaya koymaktadır. Bu bulgular, tahmin güvenilirliğini artırmak amacıyla likidite koşullarının derin öğrenme tabanlı finansal tahmin modellerine dahil edilmesinin önemini vurgulamaktadır.

Keywords

• Derin Öğrenme
• LSTM
• Likidite
• Piyasa Etkinliği
• Tahmin Hatası

The Effect of Liquidity on Deep Learning Errors

Abstract

This study investigates the impact of market liquidity and macroeconomic variables on the forecasting performance of deep learning models in financial markets. The primary objective is to forecast price movements for ten stocks listed on the BIST 30 index using a single-layer Long Short-Term Memory (LSTM) model and identify market conditions associated with prediction failures. A two-stage empirical framework is employed using daily data covering 2006–2025. First, LSTM hyperparameters are optimized through Bayesian optimization and walk-forward validation. Second, daily Absolute Percentage Error (APE) is modeled using OLS-HAC and EGARCH-X regressions to examine the effects of stock-level liquidity measures and global macroeconomic variables. Results show that the LSTM model generates well-calibrated forecasts across all ten equities, with MAPE values ranging from 3.29% to 8.40%. Trading volume and illiquidity, measured by the Amihud ratio, are positively associated with prediction errors for most stocks. At least one liquidity proxy is statistically significant for nine of the ten equities, with consistent coefficient signs where significance is observed. In contrast, macroeconomic and global indicators exhibit limited and stock-specific effects. These findings highlight the importance of incorporating liquidity conditions into deep learning-based financial forecasting models to improve forecast reliability.

Keywords

• Deep Learning
• LSTM
• Liquidity
• Market Efficiency
• Prediction Error

References

1. Alamu, O.S. and Siam, M.K. (2024). Stock price prediction and traditional models: An approach to achieve short-, medium- and long-term goals. Journal of Intelligent Learning Systems and Applications, 16, 363–383. https://doi.org/10.4236/jilsa.2024.164018
2. Alkan, S. (2024). Liquidity and market efficiency in Borsa Istanbul. Hacettepe University Journal of Economics and Administrative Sciences, 42(3), 371-384. https://doi.org/10.17065/huniibf.1388807
3. Amihud, Y. (2002). Illiquidity and stock returns: Cross-section and time-series effects. Journal of Financial Markets, 5(1), 31-56. https://doi.org/10.1016/S1386-4181(01)00024-6
4. Avcı, Ö.B. (2020). Interaction between CDS premiums and stock markets: Case of Turkey. Ömer Halisdemir University Academic Review of Economics and Administrative Sciences, 13(1), 1-8. https://doi.org/10.25287/ohuiibf.526638
5. Ballester, L. and González-Urteaga, A. (2020). Is there a connection between sovereign CDS spreads and the stock market? Evidence for European and US returns and volatilities. Mathematics, 8(10), 1667. https://doi.org/10.3390/math8101667
6. Bhattacharya, S.N., Bhattacharya, M. and Basu, S. (2019). Stock market and its liquidity: Evidence from ARDL bound testing approach in the Indian context. Cogent Economics & Finance, 7(1), https://doi.org/10.1080/23322039.2019.1586297
7. Bildirici, M., Şahin Onat, I. and Ersin, Ö.Ö. (2023). Forecasting BDI sea freight shipment cost, VIX investor sentiment and MSCI global stock market indicator indices: LSTAR-GARCH and LSTAR-APGARCH models. Mathematics, 11(5), 1242. https://doi.org/10.3390/math11051242
8. Bokka, R., Ali, M.M.R., Mouryashrith, C., Srinivas, G., Kumar, K.D.S. and Himaithika, R. (2025). An efficient hybrid model based on deep learning technique for stock price prognostication. In S. Mishra, H.K. Tripathy & J.R. Mohanty (Eds.), 2025 international conference on advancements in smart, secure and intelligent computing (ASSIC) (pp. 1-5). IEEE Computer Society. https://doi.org/10.1109/ASSIC64892.2025.11158690

9. Broto, C. and Lamas, M. (2016). Measuring market liquidity in US fixed income markets: A new synthetic indicator. The Spanish Review of Financial Economics, 14(1), 15-22. https://doi.org/10.1016/j.srfe.2016.01.001
10. Caetano, M.A.L. and Yoneyama, T. (2011). A model for the evaluation of systemic risk in stock markets. Physica A: Statistical Mechanics and its Applications, 390(12), 2368-2374. https://doi.org/10.1016/j.physa.2011.02.034
11. Chen, W., Hussain, W., Cauteruccio, F. and Zhang, X. (2024). Deep learning for financial time series prediction: A state-of-the-art review of standalone and hybrid models. Computer Modeling in Engineering & Sciences, 139(1), 187-224. https://doi.org/10.32604/cmes.2023.031388
12. Chordia, T., Roll, R. and Subrahmanyam, A. (2008). Liquidity and market efficiency. Journal of Financial Economics, 87(2), 249-268. https://doi.org/10.1016/j.jfineco.2007.03.005
13. Cont, R. and Minca, A. (2016). Credit default swaps and systemic risk. Annals of Operations Research, 247(2), 523-547. https://doi.org/10.1007/s10479-015-1857-x
14. Corwin, S.A. and Schultz, P. (2012). A simple way to estimate bid‐ask spreads from daily high and low prices. The Journal of Finance, 67(2), 719-760. https://doi.org/10.1111/j.1540-6261.2012.01729.x
15. Demir, C. (2019). Macroeconomic determinants of stock market fluctuations: The case of BIST-100. Economies, 7(1), 8. https://doi.org/10.3390/economies7010008
16. Dhokane, R.M. and Agarwal, S. (2024). A predictive model of the stock market using the LSTM algorithm with a combination of exponential moving average (EMA) and relative strength index (RSI) indicators. Journal of The Institution of Engineers (India): Series B, 105(5), 1145-1157. https://doi.org/10.1007/s40031-024-01037-8
17. Fama, E.F. (1970). Efficient capital markets: A review of theory and empirical work. The Journal of Finance, 25(2), 383-417. https://doi.org/10.2307/2325486
18. Fan, W., Wang, P., Wang, D., Wang, D., Zhou, Y. and Fu, Y. (2023). Dish-TS: A general paradigm for alleviating distribution shift in time series forecasting. Proceedings of the AAAI Conference on Artificial Intelligence, 37(6), 7522–7529. https://doi.org/10.1609/aaai.v37i6.25914
19. Febrianti, V.D. and Saadah, S. (2023). Stock liquidity and stock returns: The moderating role of financial constraints. Journal of Accounting and Investment, 24(2), 292-305. https://doi.org/10.18196/jai.v24i2.16483
20. Fischer, T. and Krauss, C. (2018). Deep learning with long short-term memory networks for financial market predictions. European Journal of Operational Research, 270(2), 654–669. https://doi.org/10.1016/j.ejor.2017.11.054
21. Gifty, A. and Li, Y. (2024). A comparative analysis of LSTM, ARIMA, XGBoost algorithms in predicting stock price direction. Engineering and Technology Journal, 9(8), 4978-4986. https://doi.org/10.47191/etj/v9i08.50
22. Gorde, P.S. and Borkar, S.N. (2024). Comparative analysis of linear regression, random forest regressor and LSTM for stock price prediction. In IEEE (Ed.), 2024 8th international conference on computing, communication, control and automation (ICCUBEA) (pp. 1-5). IEEE. https://doi.org/10.1109/ICCUBEA61740.2024.10775094
23. İlhan, A. and Akdeniz, C. (2020). The impact of macroeconomic variables on the stock market in the time of Covid-19: The case of Turkey. Journal of Research in Economics, Politics & Finance, 5(3), 893-912. https://doi.org/10.30784/epfad.810630
24. Kumar, A., Alsadoon, A., Prasad, P.W.C., Abdullah, S., Rashid, T.A., Pham, D.T.H. and Nguyen, T.Q.V. (2021). Generative adversarial network (GAN) and enhanced root mean square error (ERMSE): Deep learning for stock price movement prediction. Multimedia Tools and Applications, 81(3), 3995-4013. https://doi.org/10.48550/arXiv.2112.03946
25. Lahboub, K. and Benali, M. (2024). Assessing the predictive power of transformers, ARIMA, and LSTM in forecasting stock prices of Moroccan credit companies. Journal of Risk and Financial Management, 17(7), 293. https://doi.org/10.3390/jrfm17070293
26. Latif, S., Aslam, F., Ferreira, P. and Iqbal, S. (2025). Integrating macroeconomic and technical indicators into forecasting the stock market: A data-driven approach. Economies, 13(1), 6. https://doi.org/10.3390/economies13010006
27. Lee, H., Jeon, J., Hong, J. and Kang, U. (2025). Mitigating distribution shift in stock price data via return-volatility normalization for accurate prediction. In Proceedings of the 34th ACM International Conference on Information and Knowledge Management (pp. 1458–1467). https://doi.org/10.48550/arXiv.2508.20108
28. Li, A.W. and Bastos, G.S. (2020). Stock market forecasting using deep learning and technical analysis: A systematic review. IEEE Access, 8, 185232-185242. https://doi.org/10.1109/ACCESS.2020.3030226
29. Lundberg, S.M. and Lee, S.I. (2017). A unified approach to interpreting model predictions. In I. Guyon, U. Von Luxburg, S. Bengio, H. Wallach, R. Fergus, S. Vishwanathan and R. Garnett (Eds.), Advances in neural information processing systems 30 (pp. 4765-4774). Papers presented at the 31st Annual Conference on Neural Information Processing Systems. Long Beach, California.
30. Makarov, I. and Papanikolaou, D. (2008). Sources of systematic risk (SSRN Working Paper No. 968229). https://doi.org/10.2139/ssrn.968229
31. Mattera, R., Di Sciorio, F. and Trinidad-Segovia, J.E. (2022). A composite index for measuring stock market inefficiency. Complexity, 2022(1), 9838850. https://doi.org/10.1155/2022/9838850
32. Ooi, K.L. (2026). Analyst forecasting and market liquidity dynamics: A comparative analysis of U.S. and Chinese markets pre and post COVID-19. Journal of General Management, 51(4), 485–496. https://doi.org/10.1177/03063070231225603
33. Ribeiro, M.T., Singh, S. and Guestrin, C. (2016). "Why should I trust you?": Explaining the predictions of any classifier. In Association for Computing Machinery (Ed.), Proceedings of the 22nd ACM SIGKDD international conference on knowledge discovery and data mining (pp. 1135-1144). https://doi.org/10.1145/2939672.2939778
34. Saberironaghi, M., Ren, J. and Saberironaghi, A. (2025). Stock market prediction using machine learning and deep learning techniques: A review. AppliedMath, 5(3), 76. https://doi.org/10.3390/appliedmath5030076
35. Sonkavde, G., Dharrao, D.S., Bongale, A.M., Deokate, S.T., Doreswamy, D. and Bhat, S.K. (2023). Forecasting stock market prices using machine learning and deep learning models: A systematic review, performance analysis and discussion of implications. International Journal of Financial Studies, 11(3), 94. https://doi.org/10.3390/ijfs11030094
36. Tekin, B. and Hatipoğlu, M. (2017). The effects of the VIX index, exchange rate, and oil prices on the BIST 100 index: A quantile regression approach. Journal of Ordu University Institute of Social Sciences, 7(3), 627-634. https://izlik.org/JA89YH32ML
37. Wang, Z., Kim, S. and Joe, I. (2023). An improved LSTM-Based failure classification model for financial companies using natural language processing. Applied Sciences, 13(13), 7884. https://doi.org/10.3390/app13137884
38. Wei, Y. (2021). Absolute value constraint: The reason for invalid performance evaluation results of neural network models for stock price prediction. arXiv preprint arXiv:2101.10942. https://doi.org/10.48550/arXiv.2101.10942
39. Wu, R. and Qin, Z. (2021). Assessing market efficiency and liquidity: Evidence from China's emissions trading scheme pilots. Science of the Total Environment, 769, https://doi.org/10.1016/j.scitotenv.2020.144707
40. Zaznov, I., Kunkel, J., Dufour, A. and Badii, A. (2022). Predicting stock price changes based on the limit order book: A survey. Mathematics, 10(8), 1234. https://doi.org/10.3390/math10081234