The Relationship Of Media Information Content To Stock Returns In Chinese Stock Market

With the vigorous development of China’s stock market, investors’ attentionof media information is increasing, while the effect of media information to theattention of investors and investment strategy is growing as well. Individualinvestors are constrained to their force and cognitive ability, and they can onlyselectively handle part of all the information they have gained. So, compared toother information, the more authoritative and credible media informationbecomes the retail investors’ first choice, and then affect stock returns. Althoughpeople agree with the effect of media information on the behavior of investors aswell as the share prices of listed companies and the impact of income,comprehensive research on how media information affects investors and stockgains has just begun in our country. While the research on studying the impact onstock returns of listed companies based on the analysis of text orientation ofmedia information is even fewer. So currently, more comprehensive and in-depthresearch in this area is in great need.Firstly, using the Affective Tendency Analyzing method, this papercategorizes media information into positive and negative articles. HMSAclassification algorithm proposed by this paper is used which is also comparedwith K Nearest Neighbor algorithm, maximum entropy classification algorithmand SVM classification algorithm. The experimental results show that HMSAalgorithm performs best with accurate rate79.71%, recall rate78.00%andF-measure78.85%.After text orientation classification, the mechanism of how mediainformation content affects stock returns is verified, that is, media informationfirst impacts on the investment strategies of investors, thereby affects stock price,trading volume and yield. Through analysis, we find positive media informationcan effectively predict rise of stock yields and subsequent decrease, whilenegative media information can predict decline of stock yields and subsequentrise.On the basis of the results of the mechanism mentioned above, profit modelsof investment strategies based on the media information text orientationindicators is also developed. Since media information significantly affects stockreturns, Zero portfolio investment strategy models that with and without theindicator of media information text are considered and built. By comparison, it is found that the introduction of the indicator gets a higher return on income,increases profit margins of investors, and provides useful help to theoretical studyand practical investment.