Research On Impact Of Internet News On Stock Price Based On Event Study Methodology

As one of the main source for investors to get listed companies’ information,Internet financial news should have some impact on the fluctuation of stock price.Due to lack of scientific effective methods,existing researches often fail to quantitatively analyze text information.Previously,researchers would simply use news amount,news headlines and so on to do analysis,which would cause low prediction accuracy thus weak practical value.In this process we would need to analyze the influence of Internet news and stock prices systematically which is very important both in theory and apply fields.This article combines the Natural Language Processing Technique in Computer field and Analysis Methods in Econometrics,unites text information of news and quantitative index of stock to analyze text information of Internet news and Chinese stock market from micro perspective.From natural language processing technique perspective,text mining technology is used to turn unstructured text news into structured data.Then text classification technique based on SVM algorithm is used to divide news sample into five different categories and discuss each kind’s influence over stock price.In quantitative analysis,event study methodology is adopted to implement empirical research for detecting and comparing their significance and implement analysis.Through empirical analysis we can get following conclusions;(1)5 different kinds of Internet news reports would all significantly influence listed companies’stock price;(2)policy support category,mergers and acquisitions category,refinancing category and profitability category generate positive influence on stock price;punishment category on the contrary has negative influence on stock price;(3)investors are more sensitive to policy support,mergers and acquisitions news and lack attention on profitability category news;(4)policy support category,mergers and acquisitions category and refinancing category news have the problem of information leaking;(5)profitability category news has longest influence period on stock price while punishment category news would begin new round of influence as time passes;abnormal yield rate of mergers and acquisitions category,refinancing category news would drop down rapidly in the short time of official releasing and back to normal figure;No matter on the research content or on the research method,there is many innovations in this task.(1)Technically,improving efficiency by using automatic acquiring data technique.In the process of handling text message,transforming the unstructured text form to the structured text form,as well as classifying the news text by using classification technique and observing the impact and influence on stock price from different classes;(2)On the research content,this article did the study on the relationship of news and stock price from a new angle.Unlike past literature which simply forecast the relationship between them,this article hammered at how the different kinds of news influence the stock price,each significance level and the persistent period.In general,this article can make up the information asymmetry of stock market effectively,so there is important reference significance for investors and regulators.This article can provide theoretical basis for individual investor to do better decisions and provide theory reference for regulators to set out better rules and policy for protecting investors’ interests.It can help financial intelligence technology apply to practice more effectively.