Multi-factor Stock Selection Model And Improvement Of The Model Based On Investment Strategy Of Sentiment Index

So far,China's.securities market has developed for more than 30 years.With the improvement of laws,regulations and related systems,China's securities market has shown strong vitality.The investment enthusiasm and the market participation of ordinary people is rising.What's more,the proportion of individual investors is high all the year round,which has become a major feature of China's securities market.And the requirements for investment knowledge are also increasing.As an investment method combining mathematics,computer science and finance,quantitative investment can mine data through computer and programming language.It can also process data and develop investment strategy through mathematics and finance.Therefore,quantitative investment is a very good investment tool for individual investors.At present,compared with foreign countries,the development process of quantitative investment in China is still in its infancy.At present,scholars or research institutions focus on effective factor mining based on fundamental analysis or high-frequency trading based on technical analysis,while few studies on quantitative investment will involve behavioral finance.In view of the high proportion of individual investors in China and the fact that investors' psychology and behavior have become factors that can't be underestimated in China's stock market,this paper fits investors' sentiment as a comprehensive index and combines it with multi-factor stock selection model.From the perspective of fundamental stock selection,it uses sentiment index to guide investment activities,so as to study the effectiveness and applicability of the investment strategy based on sentiment index in the Chinese market.The empirical research of this paper is di vided into two parts.The first is to construct a multi-factor stock selection model.This paper takes all the constituent stocks of the Shanghai Composite from January 2007 to August 2020 as the research object.The paper uses the data of stocks from January 2007 to December 2011 as the sample data to establish the multi-factor stock selection model.I select 73 candidate factors and the factor screening is completed by statistical methods such as the MAD.Then I use the forward stepwise regression algorithm to complete the stock selection model construction.Second,I began to construct the sentiment index.I select 7 sentiment indicators and make principal component analysis to solve the problem of indicator lag firstly.After constructing the first-generation sentiment index,I regress the current and previous data with the first-generation sentiment index respectively and retain the data with high correlation of sentiment index.After completing the index period screening,I finish the index construction.Compared with the return rate of the next month,it was found that the emotional impact,which is the difference between the two-period sentiment index,is more obvious with the return rate.And the conclusion passes the Granger causality test.Correlation analysis,LR criterion and other results show that both lagging first-order and lagging second-order emotional shocks can have an impact on the rate of return and I use them to draw an investment strategy based on the sentiment index.Then combined the multi-factor stock selection model with the investment strategy,and compared the test results before and after adding the investment strategy.I find that the multi-factor stock selection model with the investment strategy performs better,which proves that the strategy of sentiment index can improve the result of the model.