Analysis and Forecast of Stock Index based on ARIMA-SVR Model

Qingqing Lin

《World Scientific Research Journal》 6卷11期 (2020/11) Pp. 61-65

https://doi.org/10.6911/WSRJ.202011_6(11).0009

摘要

The main purpose of this paper is to establish ARIMA model and SVR model to analyze and forecast the stock price composite index. This paper takes CSI 300 Index as an example and takes the closing price of T+1 day as the main research variable. Firstly, the linear part of the stock index is predicted by using the traditional ARIMA model. Then, the SVR model is constructed to deal with the nonlinear factors of the stock index by considering the variables such as the closing price and trading volume on t day. Considering the shortcomings of the single model, the first mock exam is to build two models and compare the prediction results with the ARIMA-SVR models. The results show that ARIMA-SVR model improves the prediction accuracy of the model to a certain extent, and can accurately predict the trend and fluctuation of CSI 300 Index in the short term, and can provide certain guidance for investors' investment decisions.

關鍵字

Stock index ； ARIMA model ； SVR model ； kernel function

參考文獻

T. B. Trafalis and H. Ince: Support vector machine for regression and applications to financial forecasting, Proceedings of the IEEE-INNS-ENNS International Joint Conference on Neural Networks, vol. 6 (2000) p. 348-353.

Chenxi Zhang, Yanping Zhang, et al. Stock forecasting based on Support Vector Machine, Computer Technology and Development, Vol. 16 (2006) No. 3, p. 35-37.

Wei Zhang: Research on stock index prediction based on linear and nonlinear Support Vector Machine combination model (MS. , Harbin Institute of technology, China 2017).

Lian Cheng: Research on personal credit evaluation method of internet finance based on Support Vector Machine (MS. , Zhejiang University of Finance and Economics, China 2017).

Yu Zhao: Air Quality Index prediction based on ARIMA and SVR combination model (MS. , Tianjin University of Commerce, China 2019).