Stock Price Volatility Prediction in Financial Big Data on XGBoost and ARIMA Models

Stock financial market is a typical incomplete information economy. Price volatility is an important characteristic of the stock market, it reflects the uncertainty of the market, but also reflects the degree of market activity. With the continuous development of digitalization and BD (Big Data) technology in the financial industry, stock price volatility prediction and analysis can help investors better understand market dynamics and formulate more scientific investment strategies. XGBoost (eXtreme Gradient Boosting) and ARIMA (Autoregressive Integrated Moving Average) models have powerful visualization capabilities and real-time performance, enabling rapid and accurate acquisition of market data. Therefore, this paper uses these two ways to describe the stock price volatility, aiming at improving the ability of stock price prediction. This paper mainly uses the experimental investigation method, case verification and comparison method, through the design of support vector machine-assisted XGBoost and ARIMA models, to get the stock market of Shenzhen and Shanghai 300 index, compare the prediction level of various models, and get relevant suggestions. Experimental results show that the recall rate of XGBoost model with support vector machine is 0.95, and the accuracy is 0.98. Therefore, adding intelligent algorithms can effectively improve the accuracy of stock price volatility prediction.