پیش‌بینی نوسانات شاخص کل بورس اوراق بهادار تهران با استفاده از شبکه عصبی مصنوعی و مدل‌های GARCH: یک مطالعه مقایسه‌ای

بیاضی, سجاد

doi:10.25843/JIFM.2025.8563.21287

پیش‌بینی نوسانات شاخص کل بورس اوراق بهادار تهران با استفاده از شبکه عصبی مصنوعی و مدل‌های GARCH: یک مطالعه مقایسه‌ای

نوع مقاله : مقاله پژوهشی

نویسنده

سجاد بیاضی

دانشجوی دکتری اقتصاد، دانشگاه علامه طباطبایی، تهران، ایران

10.25843/JIFM.2025.8563.21287

چکیده

نوسانات قیمت‌ها در بازار سرمایه یکی از ارکان اصلی تصمیم‌گیری سرمایه‌گذاری و مدیریت ریسک است. با توجه به ویژگی‌های غیرخطی و ناهمسانی واریانس شرطی در سری‌های زمانی مالی، این پژوهش قدرت پیش‌بینی مدل‌های اقتصادسنجی مالی خانواده GARCH و شبکه عصبی مصنوعی پرسپترون چندلایه را در تخمین نوسانات شاخص کل بورس اوراق بهادار تهران مقایسه می‌کند. از داده‌های روزانه شاخص کل بورس تهران در بازه فروردین ۱۳۹۳ تا اسفند ۱۴۰۲ استفاده شد. پس از محاسبه بازده لگاریتمی روزانه، مانایی و وجود اثرات ARCH بررسی و مدل‌های GARCH(1,1)، EGARCH(1,1) و GJR-GARCH(1,1) برازش شدند. سپس شبکه عصبی MLP با ورودی‌های بازده‌های باوقفه، حجم معاملات و نوسان شرطی استخراج‌شده از EGARCH طراحی و با الگوریتم لونبرگ–مارکوات آموزش داده شد. عملکرد پیش‌بینی خارج از نمونه با معیارهای RMSE، MAE، MAPE و QLIKE و آزمون دیبولد–ماریانو مقایسه گردید. آزمون‌ها مانایی بازده‌ها و وجود اثرات ARCH را تأیید کردند. ضریب اثر اهرمی در هر دو مدل نامتقارن در سطح ۱٪ معنی‌دار بود و EGARCH بر پایه AIC و BIC بهترین برازش را در میان مدل‌های خانواده GARCH داشت. در پیش‌بینی خارج از نمونه (۴۹۰ مشاهده)، رتبه‌بندی مدل‌ها از دقیق‌ترین به ضعیف‌ترین به‌صورت ANN-MLP+EGARCH، ANN-MLP، EGARCH، GJR-GARCH و GARCH(1,1) بود و آزمون دیبولد–ماریانو برتری هر دو شبکه را بر EGARCH معنی‌دار نشان داد. شبکه عصبی، به‌ویژه در ترکیب با نوسان شرطی EGARCH، دقت پیش‌بینی نوسان را نسبت به مدل‌های GARCH منفرد به‌طور معنی‌داری بهبود داد؛ از این رو ابزارهای ترکیبی هوش مصنوعی و GARCH می‌توانند در تنظیم حد زیان و مدیریت ریسک در بورس تهران سودمند باشند، هرچند تفسیر این برتری به تعریف پراکسی نوسان وابسته است.

کلیدواژه‌ها

پیش‌بینی نوسانات

بورس اوراق بهادار تهران

مدل‌های خانواده GARCH

شبکه عصبی مصنوعی

اقتصادسنجی مالی

موضوعات

مالی-بازارهای مالی

عنوان مقاله English

Forecasting Volatility of the Tehran Stock Exchange Total Index Using Artificial Neural Networks and GARCH Models: A Comparative Study

نویسنده English

Sajjad Bayazi

PhD student in Economics, Allameh Tabatabaei University, Tehran, Iran

چکیده English

Price volatility in capital markets is a fundamental element of investment decision-making and risk management. Given the nonlinear characteristics and conditional heteroskedasticity inherent in financial time series, this study compares the forecasting performance of GARCH-family econometric models and the Multilayer Perceptron (MLP) artificial neural network in estimating the volatility of the Tehran Stock Exchange (TSE) Total Index. Daily data for the TSE Total Index spanning from March 2014 (Farvardin 1393) to March 2024 (Esfand 1402) were utilized. Following the calculation of daily log returns, stationarity and the presence of ARCH effects were assessed, and GARCH(1,1), EGARCH(1,1), and GJR-GARCH(1,1) models were fitted. Subsequently, an MLP neural network was designed using inputs comprising lagged returns, trading volume, and conditional volatility derived from the EGARCH model; the network was trained using the Levenberg-Marquardt algorithm. Out-of-sample forecasting performance was evaluated using RMSE, MAE, MAPE, and QLIKE metrics, alongside the Diebold-Mariano test. Statistical tests confirmed the stationarity of returns and the presence of ARCH effects. The leverage effect coefficient was significant at the 1% level in both asymmetric models, and the EGARCH model demonstrated the best fit among the GARCH-family models based on AIC and BIC criteria. In the out-of-sample forecast (490 observations), the ranking of models from most accurate to least accurate was ANN-MLP+EGARCH, ANN-MLP, EGARCH, GJR-GARCH, and GARCH(1,1); the Diebold-Mariano test indicated that the superiority of both neural network models over EGARCH was statistically significant. Neural networks—particularly when combined with EGARCH conditional volatility—significantly improved volatility forecasting accuracy compared to standalone GARCH models; thus, hybrid AI-GARCH tools can be beneficial for setting stop-loss limits and managing risk in the Tehran Stock Exchange, although the interpretation of this superiority depends on the definition of the volatility proxy.

کلیدواژه‌ها English

Volatility forecasting

Tehran Stock Exchange

GARCH family models

Artificial neural network

Financial econometrics

Aminimehr, A., et al. (2020). The role of feature engineering in prediction of Tehran stock exchange index based on LSTM. Iranian Journal of Economic Studies, 9(2), 527–548.

Adibi, M. A., & Pourghader Chobar, A. (2025). Decision intelligence to enhance bank profitability through customer promotion path design. Journal of Systems Thinking in Practice, 4(1), 70-82.

Adibi, M. A., Pourghader Chobar, A., & Mortazi, S. (2025). A mathematical model for two-echelon open vehicle routing problem with time-windows and mobile satellites based on customer satisfaction. Opsearch, 1-34.

Bildirici, M., & Ersin, Ö. Ö. (2009). Improving forecasts of GARCH family models with the artificial neural networks: An application to the daily returns in Istanbul Stock Exchange. Expert Systems with Applications, 36(4), 7355–7362.

Bollerslev, T. (1986). Generalized autoregressive conditional heteroskedasticity. Journal of Econometrics, 31(3), 307–327.

Boroumand, O., & Doaei, M. (2024). Developing a stock market prediction model by deep learning algorithms. Journal of Information Technology Management, 16(3), 115–131. https://doi.org/10.22059/jitm.2024.357529.3311

Bucci, A. (2020). Realized volatility forecasting with neural networks. Journal of Financial Econometrics, 18(3), 502–531.

Daneshvar, A., Olfat, M., Pourghader Chobar, A., & Yadegari, E. (2025). Improving the performance of direct advertising campaigns by applying artificial intelligence techniques. Corporate Communications: An International Journal.

Dickey, D. A., & Fuller, W. A. (1979). Distribution of the estimators for autoregressive time series with a unit root. Journal of the American Statistical Association, 74(366), 427–431.

Diebold, F. X., & Mariano, R. S. (1995). Comparing predictive accuracy. Journal of Business & Economic Statistics, 13(3), 253–263.

Engle, R. F. (1982). Autoregressive conditional heteroscedasticity with estimates of the variance of United Kingdom inflation. Econometrica, 50(4), 987–1007.

Gebreslase, T. T., Mutua, K., Fissuh, Y. H., & Girmay, T. G. (2025). A hybrid multi-GARCH transformer model for stock price volatility forecasting. Communications in Mathematical Biology and Neuroscience.

Glosten, L. R., Jagannathan, R., & Runkle, D. E. (1993). On the relation between the expected value and the volatility of the nominal excess return on stocks. The Journal of Finance, 48(5), 1779–1801.

Hagan, M. T., & Menhaj, M. B. (1994). Training feedforward networks with the Marquardt algorithm. IEEE Transactions on Neural Networks, 5(6), 989–993.

Hansen, P. R., & Lunde, A. (2005). A forecast comparison of volatility models: Does anything beat a GARCH(1,1)? Journal of Applied Econometrics, 20(7), 873–889.

Hornik, K., Stinchcombe, M., & White, H. (1989). Multilayer feedforward networks are universal approximators. Neural Networks, 2(5), 359–366.

Hu, Y., Ni, J., & Wen, L. (2020). A hybrid deep learning approach by integrating LSTM-ANN networks with GARCH model for copper price volatility prediction. Physica A: Statistical Mechanics and its Applications, 557, 124907.

Jarque, C. M., & Bera, A. K. (1980). Efficient tests for normality, homoscedasticity and serial independence of regression residuals. Economics Letters, 6(3), 255–259.

Keihani, H., & Keihani, H. (2025). Transitioning corporate models and processes towards sustainable practices and adopting a circular economy approach. Canadian Journal of Business, Economics and Health Managemen, 8(2), 31-46.

Keihani, H. (2021). Macro Economic Factors Influencing the Indian Stock Market with Special Reference to Bombay Stock Market. Shodh Sarita, 8(29), 263-268.

Kim, H. Y., & Won, C. H. (2018). Forecasting the volatility of stock price index: A hybrid model integrating LSTM with multiple GARCH-type models. Expert Systems with Applications, 103, 25–37.

Mahajan, V., Thakan, S., & Malik, A. (2022). Modeling and forecasting the volatility of NIFTY 50 using GARCH and RNN models. Economies, 10(5), 102. https://doi.org/10.3390/economies10050102

Marquardt, D. W. (1963). An algorithm for least-squares estimation of nonlinear parameters. Journal of the Society for Industrial and Applied Mathematics, 11(2), 431–441.

Motevalli, S. H., Chobar, A. P., Ebrahimi, M., & Alamiparvin, R. (2024). A Fuzzy Multi-objective Optimization Model in Sustainable Supply Chain Network Design Considering Financial Flow. Fuzzy Optimization and Modeling Journal (FOMJ), 5(1), 27-45.

Nabipour, M., Nayyeri, P., Jabani, H., Mosavi, A., Salwana, E., & Shahab, S. (2020). Deep learning for stock market prediction. Entropy, 22(8), 840. https://doi.org/10.3390/e22080840

Nelson, D. B. (1991). Conditional heteroskedasticity in asset returns: A new approach. Econometrica, 59(2), 347–370.

Nozari, H., & Chobar, A. P. (2024). The Dimensions and Components of Marketing 5.0: Introduction to Marketing 6.0. In Advanced Businesses in Industry 6.0 (pp. 75-86). IGI Global Scientific Publishing.

Patton, A. J. (2011). Volatility forecast comparison using imperfect volatility proxies. Journal of Econometrics, 160(1), 246–256.

Phillips, P. C. B., & Perron, P. (1988). Testing for a unit root in time series regression. Biometrika, 75(2), 335–346.

Poon, S.-H., & Granger, C. W. J. (2003). Forecasting volatility in financial markets: A review. Journal of Economic Literature, 41(2), 478–539.

Roszyk, N., & Ślepaczuk, R. (2024). The hybrid forecast of S&P 500 volatility ensembled from VIX, GARCH and LSTM models. arXiv preprint arXiv:2407.16780.

Sadon, A. N., Ismail, S., Khamis, A., & Tariq, M. U. (2024). Heteroscedasticity effects as component to future stock market predictions using RNN-based models. PLOS ONE, 19(5), e0297641.

Shakerian, M., & Keihani, H. (2026). AI-Driven Economic Forecasting for Industry 5.0 Enterprises. In Industry 5.0's Impact on Economic Innovation (pp. 35-50). IGI Global Scientific Publishing.

Tan, C. Y. (2025). Predicting stock market volatility: A comparative analysis of machine learning models and GARCH. Tunku Abdul Rahman University of Management and Technology (institutional repository).

Wang, H., Liu, C., Tran, M.-N., & Wang, C. (2025). Deep learning enhanced multivariate GARCH. arXiv preprint arXiv:2506.02796.

Xu, Z., Liechty, J., Benthall, S., Skar-Gislinge, N., & McComb, C. (2024). GARCH-informed neural networks for volatility prediction in financial markets. arXiv preprint arXiv:2410.00288.