Research On Multi-factor Stock Selection Based On Baidu Index

The capital market in China is still immature,which leads to a discrepancy between the price and true value of many securities.This difference gives investors the opportunity to obtain income from it,and also provides room for the development in quantitative trading.At present,foreign quantitative trading strategies have developed relatively mature,but domestic quantitative trading is still in the process of continuous improvement.Against the background of the normalization of the global epidemic and the "Healthy China 2030",the pharmaceutical industry is currently an industry with high investment value and high growth in the domestic market.It is not only supported by national strategy,but also by the rich profits of the industry itself.In recent years,the trend of pharmaceutical stocks have gained strong momentum,and the accumulated income has significantly outperformed the market,so it deserves to be invested.And there are sufficient sample stocks in the pharmaceutical industry,and the author is more familiar with it,so the author chooses the pharmaceutical industry as a research field.This paper takes the constituent stocks that have appeared in the CSI Health Care over the years as the stock pool,and innovatively introduces the Baidu Index factor,thus constructing the multi-factor stock selection strategies in the pharmaceutical industry.Firstly,in the factor screening stage,this paper constructs an initial factor database including Baidu index and preconditions the data of each factor.Secondly,three factors of Baidu Index are tested by Fama-Mac Beth regression,which proves the validity of mobile index and search index in Baidu Index as stock selection factors.Then,all the factors are tested by IC,IR and group test,and the redundant factors are eliminated,and the final effective factors are obtained on the basis of the results of the correlation coefficient table and the variance expansion coefficient table.In the back-test evaluation stage,three models are constructed by means of scoring method,and all of them are tested by the CSI 300 Index and the CSI Health Care Index as the benchmark.The results show that:(1)In the pharmaceutical industry,the multi-factor stock selection model is effective,and the three multi-factor stock selection models constructed in this article performed better than the CSI 300 index and the CSI Health Care Index during the back-test.(2)The mobile index and search index in Baidu Index are significantly related to the trend of stocks.Fama-Macbeth test shows that the mobile index is more effective,and IC test shows that the search index is more effective.However,the mobile index is better as a stock selection factor in accordance with the back test results.(3)Baidu index can optimize the multi-factor stock selection stock selection model built only with traditional factors.And adding the search index and mobile index to the multi-factor stock selection model based on traditional factors can bring more information to investors,so that investors beat the market to obtain excess returns.