The quantitative trading forecast model of stock market is an important research direction of the development of financial industry and computer industry in recent years.How to make maximum use of the characteristics of financial data itself to enhance the predictive power of machine learning is the focus of this paper.Since traditional machine learning methods lack the mining of intrinsic associations of feature indicators according to market characteristics and high-dimensional nonlinear data characteristics,and only generalize the data set with subjective and objective factors and mostly research more for a specific data,the application scope is limited,this paper proposes a DMD-XGBoost quantitative stock market trading prediction model based on e-commerce concept stocks.Data innovation uses 12 financial characteristics indicators of daily trading data of ecommerce platform concept stocks for the period from September 1,2019 to September 1,2021 for multi-perspective analysis.The empirical results show that the algorithm proposed in this paper can effectively identify stock trading information,and the comparison between the quantified and traditional models by establishing an evaluation index system yields an RMSE value of about 0.1178,both of which are higher than the traditional research methods,indicating that the improvement in prediction effect is achieved while expanding the application scenario.By integrating machine learning and stock market,this paper has changed the shortcomings of the traditional model in terms of explanatory power,and provided new ideas on the correlation and interaction between indicators,and improved the prediction effect,which confirms the value and feasibility of the interdisciplinary treatment of "big data+mathematics+finance". |