پیش بینی قیمت سهام در بازار سرمایه با رویکرد هوش مصنوعی

محورهای موضوعی : پژوهشنامه اقتصاد و کسب و کار

عسگر نوربخش ^{1 *} , مصطفی شایگانی ²

1 - استادیار، گروه مدیریت مالی، دانشکده مدیریت و حسابداری، دانشکدگان فارابی دانشگاه تهران، قم، ايران
2 - کارشناسی ارشد مدیریت مالی، دانشکده مدیریت و حسابداری، دانشکدگان فارابی دانشگاه تهران، قم، ايران

تاریخ دریافت : 1403/03/20 تاریخ پذیرش : 1403/05/25 تاریخ انتشار : 1403/10/14

کلید واژه: شبکه عصبی مصنوعی, حافظه کوتاه مدت ماندگار, شبکه عصبی پیچشی, پیش‌بینی قیمت,

چکیده مقاله :

هدف این پژوهش پیش‌بینی قیمت سهام با استفاده از دو نوع شبکه عصبی در بورس اوراق بهادار تهران است. شبکه‌های عصبی بازگشتی[1] عموماً در پیش‌بینی داده‌های سری زمانی توانایی خوبی دارند، اما شبکه‌ی عصبی پیچشی[2] عمدتا برای کاربردهایی چون بینایی کامپیوتر استفاده می‌شوند. برای انجام این پژوهش از زبان پایتون در ویرایشگر VS code استفاده شده است. جامعه آماری این پژوهش بورس اوراق بهادار تهران می‌باشد. حجم نمونه آماری این پژوهش شامل داده‌های سه نماد بورس اوراق بهادار تهران به شرح ایران خودرو، البرز دارو و توسعه معادن روی ایران است. در این پژوهش از هشت ویژگی قیمت در چارچوب زمانی روزانه از تاریخ 1380 تا تاریخ 1400 استفاده می‌شود که شامل بالاترین قیمت، پایین‌ترین قیمت، قیمت بسته‌شدن، قیمت باز شدن، ارزش معاملات، حجم معاملات، اختلاف قیمت بسته شدن دو روز متوالی، و بازده روزانه است. برای ارزیابی عملکرد مدل‌ها از سه معیار خطای میانگین خطای مطلق، ریشه میانگین مربعات خطا و ضریب تعیین استفاده شده است. نتایج نشان می‌دهد که مدل شبکه عصبی پیچشی توانایی پیش بینی با دقت خوبی را دارا می‌باشد. شبکه‌های عصبی بازگشتی از بهترین نوع شبکه‌ها برای پیش‌بینی قیمت هستند، اما نتایج نشان می‌دهد که شبکه عصبی پیچشی عملکرد بهتری از شبکه عصبی حافظه کوتاه مدت ماندگار داشته است. نتایج نشان می‌دهد که مدل‌های یادگیری عمیق در صورتی که در انتخاب ویژگی‌هایی (متغیرهای مستقل) که بتوانند بیشترین میزان معناداری را در تفسیر علل فراز و فرودهای قیمت در دوره‌های رونق و رکود بازار داشته باشند، قابلیت و توانایی پیش‌بینی قیمت، با دقت قابل قبول را خواهند داشت.

---

[1] . Recurrent Neural Network (RNN)

[2] . Convolutional neural network (CNNs)

چکیده انگلیسی:

The main objective of this research is to predict stock prices using two types of neural networks in the Tehran Stock Exchange. Python language in the VS Code editor has been used to conduct this research. The statistical population of this research is the Tehran Stock Exchange. The sample size of this research includes data from three symbols of the Tehran Stock Exchange, namely Iran Khodro, Alborz Darou, and Iran Zinc Mines Development. In this research, eight price features are used within the daily timeframe from 2001 to 2021, including the highest price, lowest price, closing price, opening price, transaction value, transaction volume, the difference in closing price between two consecutive days, and daily return. Three metrics of mean absolute error, root mean square error, and coefficient of determination have been used to evaluate the models' performance. The results indicate that the convolutional neural network model has the ability to predict with good accuracy. Recurrent neural networks are among the best types of networks for price prediction, but the results show that the convolutional neural network has performed better than the short-term memory neural network. The results suggest that deep learning models, when selecting features (independent variables) that can express the highest level of significance in interpreting the causes of price fluctuations during market booms and recessions, have the ability and capability to predict prices with acceptable accuracy.

منابع و مأخذ:

Vanaga, R., & Sloka, B. (2020). Financial and capital market commission financing: aspects and challenges. Journal of Logistics, Informatics and Service Science, 7(1), 17–30
Zhang, L., & Kim, H. (2020). The influence of financial service characteristics on use intention through customer satisfaction with mobile fintech. Journal of System and Management Sciences, 10(2), 82–94
Badea, L., Ionescu, V., & Guzun, A.-A. (2019). What is the causal relationship between Stoxx Europe 600 sectors? But between large firms and small firms? Economic Computation And Economic Cybernetics Studies And Research, 53(3), 5–20
Sousa, J., Montevechi, J., & Miranda, R. (2019). Economic lot-size using machine learning, parallelism, metaheuristic and simulation. Journal of Logistics, Informatics and Service Science, 18(2), 205–216
Coser, A., Maer-Matei, M. M., & Albu, C. (2019). Predictive models for loan default risk assessment. Economic Computation And Economic Cybernetics Studies And Research, 53(2), 149–165
Qiao, R. (2019). Stock prediction model based on neural network. Operations Research and Management Science, 28(10), 132–140
Jung, C., & Boyd, R. (1996). Forecasting UK stock prices. Applied Financial Economics, 6(3), 279–286
Bleesser, W., & Liicoff, P. (2005). Predicting stock returns with Bayesian vector autoregressive. Data Analysis, Machine Learning and Applications, 1, 499–506
Adebiyi, A., Adewumi, A., & Ayo, C. (2014). Stock price prediction using the ARIMA model. In Proceedings of the 2014 UKSim-AMSS 16th International Conference on Computer Modelling and Simulation. IEEE, Cambridge, UK
Zhang, C., Cheng, X., & Wang, M. (2005). An empirical research in the stock market of Shanghai by GARCH model. Operations Research and Management Science, 4, 144–146
Yang, Q., & Wang, C. (2019). A study on forecast of global stock indices based on deep LSTM neural network. Statistical Research, 36(6), 65–77
Moon, K.-S., & Kim, H. (2019). Performance of deep learning in prediction of stock market volatility. Economic Computation And Economic Cybernetics Studies And Research, 53(2), 77–92
Li, J., Pan, S., Huang, L., & Zhu, X. (2019). A machine learning based method for customer behavior prediction. Tehnicki Vjesnik-Technical Gazette, 26(6), 1670–1676
White, H. (1988). Economic prediction using neural networks: the case of IBM daily stock returns. Earth Surface Processes & Landforms, 8(5), 409–422
Zhang, G. P. (2003). Time series forecasting using a hybrid ARIMA and neural network model. Neurocomputing, 50(1), 159–175
Sun, Y., Liang, Y., & Zhang, W. (2005). Optimal partition algorithm of the RBF neural network and its application to financial time series forecasting. Neural Computing and Applications, 14, 1441–1449
Adhikari, R., & Agrawal, R. (2014). A combination of artificial neural network and random walk models for financial time series forecasting. Neural Computing and Applications, 24, 305–315
Zhang, L., Wang, F., Xu, B., Chi, W., Wang, Q., & Sun, T. (2018). Prediction of stock prices based on LM-BP neural network and the estimation of overfitting point by RDCI. Neural Computing and Applications, 30(5), 1425–1444
Hu, Y. (2018). Stock market timing model based on convolutional neural network – a case study of Shanghai composite index. Finance& Economy, 4, 71–74
Alibasic, E., Fazo, B., & Petrovic, I. (2019). A new approach to calculating electrical energy losses on power lines with a new improved three-mode method. Tehnicki Vjesnik-Technical Gazette, 26(2), 405–411
Xue, Y., Wang, C., & Miao, C. (2020). Research on financial assets transaction prediction model based on LSTM neural network. Neural Computing and Applications, 1
LeCun, Y., Bottou, L., Bengio, Y., & Haffner, P. (1998). Gradient-based learning applied to document recognition. Proceedings of the IEEE, 86(11), 2278–2324
Kim, B. S., & Kim, T. G. (2019). Cooperation of simulation and data model for performance analysis of complex systems. International Journal of Simulation Modelling, 18(4), 608–619
Qin, L., Yu, N., & Zhao, D. (2018). Applying the convolutional neural network deep learning technology to behavioural recognition in intelligent video. Tehnicki Vjesnik-Technical Gazette, 25(2), 528–535
Hochreiter, S., & Schmidhuber, J. (1997). Long short-term memory. MIT Press, 9(8), 1735–1780
Ta, V.-D., Liu, C.-M., & Tadesse, D. A. (2020). Portfolio optimization-based stock prediction using long-short term memory network in quantitative trading. Applied Sciences, 10(2), 437–457
Zarrad, O., Hajjaji, M. A., & Mansouri, M. N. (2019). Hardware implementation of hybrid wind-solar energy system for pumping water based on artificial neural network controller. Studies in Informatics and Control, 28(1), 35–44
Saric, T., Simunovic, G., Vukelic, D., Simunovic, K., & Lujic, R. (2018). Estimation of CNC grinding process parameters using different neural networks. Tehnicki Vjesnik-Technical Gazette, 25(6), 1770–1775
Gupta, N., & Jalal, A. (2019). Integration of textual cues for fine-grained image captioning using deep CNN and LSTM. Neural Computing and Applications, 12, 1–10
Yadav, A., Jha, C. K., & Sharan, A. (2020). Optimizing LSTM for time series prediction in Indian stock market. Procedia Computer Science, 167, 2091–2100
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
Petersen, N. C., Rodrigues, F., & Pereira, F. C. (2019). Multi-output bus travel time prediction with convolutional LSTM neural network. Expert Systems with Applications, 120, 426–435
Jin, Z., Yang, Y., & Liu, Y. (2020). Stock closing price prediction based on sentiment analysis and LSTM. Neural Computing and Applications, 32(13), 9713–9729
Svetlana, B., & Ioannis, T. (2019). An ensemble of LSTM neural networks for high-frequency stock market classification. Journal of Forecasting, 38(6), 600–619