A hybrid forecasting method for quantitative investments based on GA-VMD and SSA-DELM optimization

Abstract

In the modern financial market, quantitative investment is a more advanced investment method, which has been widely used in the capital market with the characteristics of simplicity, efficiency, and plasticity. Compared with other countries, China started late in the quantitative field, and the strategies are not perfect enough to be studied in depth. To study quantitative investment in-depth, this paper proposes a GA-VMD-SSA-DELM quantitative model. By building an extreme learning machine (DELM) model optimized by the sparrow algorithm (SSA), and simultaneously using a variational modal decomposition model (VMD) optimized by the genetic algorithm (GA) for data noise reduction, a quantitative investment model is constructed, and a portfolio strategy is formed. The method first performs correlation tests on the data sources, uses the daily frequency trading indicators of Guizhou Maotai, Zhifei Bio, and Yangtze River Power for the years 2018-2022 as the database, and uses SPSS to conduct Pearson correlation coefficient analysis to establish the correlation between the data. Then the GA-VMD algorithm was used for data noise reduction. Finally, the SSA-DELM algorithm is used to establish a quantitative investment model, construct a portfolio, and draw relevant conclusions.