Empirical Study on Multi-Factor Quantitative Stock Selection Model Based on Support Vector Machine

Machine learning has many applications in the financial field due to its superior data analysis ability and non-linear fitting ability. Multi-factor stock selection plays a crucial role in quantitative investment, demonstrating consistent performance across various market conditions. In order to effectively monitor the driving factors of China's stock market, this article identifies the influential factors that are effective in the A-share market. The paper uses SVM algorithm for multi-factor stock selection, with the CSI300 (China Securities Index 300) Component Stocks as the research object. The individual stock returns are predicted and the optimal investment portfolio is constructed by training the factor data of each component stock. The SVM stock selection model with a warehouse transfer frequency of 30 trading days exhibits an impressive annualized return of 15.70% based on empirical findings, which is better than the SVM model with a warehouse transfer frequency of 60 trading days and the SVM model with a warehouse transfer frequency of 90 trading days.