Pricing Theory And Empirical Study Of China GEM Stocks Based On Text-sentiment Analysis

Currently,the estimation of risk on stock market is a challenge for modern finance theory.Although it has a significant influence on investors and market regulators,but it is still under debate which risks should be concerned for pricing and cost estimation.This article starts from the analysis of the behavioral finance theory,emphasizing its application on asset pricing.Then calculate the “public sentiment” indicator,which represents investor sentiment,using data mining techniques-an interdisciplinary method.On the basis of this,I adjust the Fama-French three-factor model.Finally,I respectively test the applicability of the two models in GEM from an empirical point of view.The GEM listed companies are divided into six groups according to size and book capitalization ratio.Through statistical analysis,the author found “small firm effect” in high BM companies and “value effect” in small companies.The regression analysis showed that the extended four-factor model fit GEM better,namely,public sentiment can explain the stock return volatility.Periodic analysis showed that there is a certain “herding” on the GEM.Therefore,we use sentiment analysis techniques to quantify the public network into public sentiment indicators may provide a new thinking for the study of behavioral finance.