Stock Volatility Prediction Analysis Based on Different Models

Abstract

Stock volatility prediction is an important method for making the correct financial decision in modern stock market. Additionally, traditional method of stock indices prediction has been proved by modern scholars to be ineffective. This review focuses on the volatility prediction performance of various models, aiming to figure out the most effective volatility prediction model for modern stock market. The four models discussed in this review are ARIMA, XGBOOST, GARCH, and LSTM. Concluding the effectiveness of different models in volatility prediction can be helpful for practical applications of these models in modern stock markets, which can improve the profitability of stock investors under certain situations. Among all four models discussed in this review, LSTM model shows the best performance among all four models in stock volatility prediction. The unique machine learning ability of LSTM has great potential value in future science field research. Since this study is only a general review of stock volatility prediction models, additional research and analysis are essential in order to further conduct this topic for better accuracy and credibility.