Construction Of Multi-factor Stock Selection Strategy In Textile And Apparel Industry

Since the birth of the capital market,investors and researchers have been focusing on finding effective investment screening methods.During the period,a variety of investment analysis theories were produced.Among them,quantitative investment theory is highly applicable to mathematics because of its high degree of integration with mathematics.It is favored by scholars and investors.Multi-factor model is one of the most widely used and mature quantitative stock selection models in the field of quantitative investment.In recent years,as econometric models have gradually entered the vision of scholars,the combination of quantitative models and portfolio investment has become a new trend in the future..The textile and apparel industry is one of China's important export industries.Especially under the guidance of the new era and new ideas,China's textile and apparel industry has formed an independent and complete modern industrial system.Driven by the policy backgrounds of the “Belt and Road” and “Made in China 2025” and domestic and foreign demand,the textile and apparel industry has entered the development stage of industrial structure optimization and upgrading.Industry growth and macro policy backgrounds have highlighted textile and apparel in many aspects.The investment value of the industry.Based on the above industry background and theoretical analysis,all listed companies in China's A-share textile and apparel industry were selected as the research object(the time interval is from 2012 to 2020),and 38 were selected with reference to the "China A-Share Market Quantification Factors White Paper".After candidate factors,the traditional IC factor regression method was used to screen the candidate factors.At the same time,the LASSO regression model was used to further test the validity of the factors by adding a penalty function.The effective factors were scored by the scoring method and the size of each factor was added up The total score and ranking of each stock are obtained,and the top 5 stocks are selected as the investment portfolio.The proportion of individual stocks in the portfolio is the proportion of their stock scores to the total score of the stock portfolio.Finally,the backtesting is used to check the portfolio.Actual market performance.The backtesting results show that the effective factors selected by the traditional IC regression method are not accurate enough and the coverage is not complete.The stock selection strategy constructed based on this can not effectively diversify risks,combine poor returns and benchmark returns.There is a significant correlation between the effective factors screened by the lasso regression model and the stock return,and the stock selection strategy constructed by it has an excellent effect,which can outperform the benchmark index from both perspectives of return and risk.In addition,by continuously increasing the penalty function in the lasso regression model,we can obtain better stock selection strategies.The applicability and robustness of the combination of lasso regression model and multi-factor stock selection strategy in the textile and apparel industry were verified.