Quantitative investment model based on LightGBM algorithm
Yangyang Guo · Academic Journal of Computing & Information Science · 2022
Quantitative investment refers to the trading method through quantitative way and computer program issued orders, in order to obtain stable income. Firstly, this paper completes Pearson correlation analysis. Finally, the correlation coefficients of 50 indicators were calculated according to relevant data, and 8 indicators with the highest correlation were obtained. Secondly, according to the LightGBM model, multiple linear regression model and BP neural network model, the three prediction models are used to forecast the trading volume of the "digital economy" plate from January 4, 2022 to January 28, 2022 respectively. Thirdly, the mean absolute value error, mean square error, R and other model evaluation indexes of the three prediction models are analyzed. Finally, the accuracy of LightGBM model is better, followed by BP neural network model, and multiple linear regression is the worst.