Quantitative Investment Strategy Based On The Comovement Between Stocks

There are various comovement in the financial market,among which the comovement between different markets and between related industry sectors are widely studied,in the past research,these findings have a guiding role for investors to make long-term investment,and in the practice,researchers and investors can easily make judgments through experience.However,in the whole market,the linkages between stocks are more complex.Many potential impacts and linkages among different stocks are not easy to detect through investment experience and analysis,they rely more on market changes and external environment,leading them difficult to make judgments.This paper based on the comovement between stocks,and took the CSI 300 index as the market representative,combined the stock selection method on Support Vector Regression(SVR)and the timing trading method on LASSO algorithm,constructed a quantitative investment strategy,and retested during the time interval from May 1,2016 to May 5,2022,evaluated the prediction performance,income and process performance of the strategy through the confusion matrix,yield,maximum withdrawal rate,Sharpe Ratio,Information Ratio.The findings are as follows:(1)Compared with the CSI 300 index,the stock selection strategy had a better performance of returns in the holding state,but had no advantage in anti-risk ability.(2)The stock price trend prediction based on the linkage effect between stocks had great prediction performance,and compared with the holding strategy,with the support of the timing trading strategy,the asset portfolio achieved higher excess returns and significantly improved its anti-risk ability.(3)In the rising period of market,the timing trading strategy can track gains,but there will be a slight loss of earnings;In the falling period,the strategy can better control loss and reduce risks;While in the shock period,the strategy can effectively reduce the profit and loss volatility of the asset portfolio.Compared with the previous research on qualitative analysis and strategy construction based on fixed factors or comovement between different markets,this paper combined the sparsity,mutation,timeliness and hysteresis of the comovement between stocks to construct time-selective trading strategies,selected the LASSO model with strong adaptability,and applied the unknown comovement between individual stocks,providing a new method for the quantitative investment strategies.Besides,the strategy and algorithm methods of this paper also had good applicability and flexibility.This research also has a large extension space based on different object and purpose,we can expand it in the field of different frequency of trading strategy,strategy performance research in different markets,financial interpretative research behind comovement,and so on.