Visualization Analysis of Stock Data and Intelligent Time Series Stock Price Prediction Based on Extreme Gradient Boosting

This paper discusses the stock market with a complex nonlinear system and establishes a time series prediction model based on XGBoost by using the idea and theoretical basis of chaos theory. This paper analyzes the daily trend, weekly trend, and monthly trend of the three stocks and forecasts the future rise and fall of the price trend obtained from the data of different stocks. Through the rise and fall data to verify and analyze the cyclical changes of the stock and summarize the law. In the end, from the visualization results of the model, the trend and closing price fitted by the model are good. Therefore, according to the comparative analysis, it can be concluded that the XGBoost model can basically predict the general trend of stock prices. Also, not all stocks have a price reversal cycle phenomenon, and the influencing factors and variables of the stock rise and fall are chaotic. In the future, the turnover rate can reflect the liquidity of the stock in the market. Also, our line graph makes it more fitting. The accuracy of the model can be higher. The code is available at https://github.com/Fickle519Istockforecastingmodelbasedon_XGboost.git.