Forecasting Method of Stock Market Volatility Based on Multidimensional Data Fusion

Abstract

The volatility of the stock market is related to the vital interests of stockholders and is essential for maintaining a stable financial environment. Through the analysis of data changes, excellent professional traders can extract information about the direction of stock changes, whether it is worth investing, and long-term or short-term trading. This article aims to study the forecasting methods of stock market volatility, by integrating multiparty data, in-depth analysis of the direction of data changes, predicting the price changes of the stock market, and better guiding stockholders’ investment. This paper proposes a multisource data fusion method to analyze the stock market price changes and find the best risk prediction method. The experimental results in this paper show that multisource data fusion can better help the stock market predict stock changes and reduce financial investment risks by 20%. Comparing the obtained prediction results with the real data, the MSE predicted by the ARIMA model is calculated to be 2.35. It provides a new idea for effectively analyzing nonstationary time series data with complex trend fusion characteristics by rationally screening feature signals and trend signals and modeling probability distribution.