Application of Gradient Boosting Algorithm in Investment Trends Forecast

In recent years, the pulling role of investment in economic growth is increasingly prominent compared to consumption and exports due to the impact of the epidemic. In order to actively cope with the triple pressure of demand contraction, supply shock, and weakening expectations, as well as promote economic stabilization and recovery, it is essential to strengthen the systematic analysis and advance prediction of the investment trends. This study uses LightGBM to implement the gradient boosting algorithm to construct a model of the distribution pattern of investment to forecast the investment trends. The model is verified by using the data of intended investment projects in Zhejiang Province for two years, 2021 and 2022, and the error is within 15%. According to the forecast, the growth rate of fixed asset investment in Zhejiang Province in the first half of 2023 is about 9.5%- 10.4%.