The Stock Index Rise And Fall Prediction Program Based On The LightGBM Algorithm

The prediction of financial time series has always been the focus of investors and scholars,but the stock market is affected by many complex factors,so it is difficult to predict.With the development of computer technology,more and more machine learning methods are applied to the prediction of financial time series,especially in the prediction of the rise and fall of stock price index.Based on the lightgbm algorithm,this paper constructs a prediction model for the rise and fall of the stock index,and forecasts the rise and fall of the CSI 300 index and CSI 500 index.12 technical indicators are selected as the input characteristics,and the rise and fall of the next day are the output variables.At the same time,in order to optimize the prediction effect of the model,the LR lightgbm model is constructed to predict the two stock indexes,and the grid search method is used to optimize the parameters of the algorithm to compare the prediction effect of the above two models.This paper selects the trading day data from January 1,2012 to December 31,2019 for empirical research.The results show that the above prediction model has good applicability to the prediction of the rise and fall of the stock index,the prediction effect for the CSI 300 index is good,and the logical regression is helpful to improve the prediction accuracy of the lightgbm model.Then,based on the prediction results of lightgbm model and LR lightgbm model,a quantitative timing strategy is constructed,and Jiashi Hushen 300 etf is taken as the trading partner for back testing.From the aspect of risk and return,we compare the two models synthetically,and find that the LR lightgbm model is better,and it can be used as a reference for investors.