The Research On Dynamic Multi-factor Stock Selection Model Based On Timing Method

In recent years,quantitative investment research has developed rapidly.From the point of investment practice,the short-term volatility of the market is increasing,and portfolio returns are often affected by macro fundamentals,market performance and other factors.Investors pay more and more attention to the allocation of selection factors,highlighting the necessity of factor timing.Multi-factor alpha model is the core of quantitative investment system.So far,there is no unified solution to choose factors and how to allocate the weight of factors.Therefore,according to the existing literature and the actual situation of Chinese stock market,this paper selects and synthesizes 8 style factors from 50 factors through the analysis of single factor validity from the two angles of factor selection and weight allocation of multi-factor model.In the aspect of weighting method,the existing HKQ factor timing model is improved,and a more stable and effective stock selection model suitable for domestic stock market is established through dynamic allocation of factor weights.Then,in the A-share market,three different multi-factor stock selection models are constructed,which are static equal weight,IC dynamic weighting model and improved dynamic weighting model according to different factor weighting methods.The model is debugged and perfected,and the performance of each model is compared and analyzed.The results show that the performance of synthetic factor in risk and profit is significantly higher than that of single factor.The overall performance of the three models is as follows: static equal weighting < IC dynamic weighting < improved dynamic weighting in 2012.09.03~2018.03.30.The three yields were 25.23%?32.98%?39.88%,The three models can produce excess returns compared with the CSI 300 index,the risk coefficient does not decrease.However,the risk of hedging the dynamic model has a significant decline.The improved dynamic weighting model has the most significant increase in income,and Sharp ratio is the largest.It shows that the dynamic model can make up for the deficiency of the static model when the market fluctuates greatly,and the dynamic model considering the influence of exogenous variables is better than the IC dynamic model.