Font Size: a A A

A Study On The Investor Sentiment Drving Strategy Model Based On Text Mining

Posted on:2018-10-29Degree:MasterType:Thesis
Country:ChinaCandidate:W K LiFull Text:PDF
GTID:2359330542969838Subject:Management Science and Engineering
Abstract/Summary:PDF Full Text Request
With the rapid development of the Internet,more and more investors voice their opinions in the network platform such as BBS and microblog.And its impact on financial markets has become important.As an significant part of behavioral finance,investor sentiment has long been a research hotspot.Now how to measure investor sentiment on the network has also become a current research hotspot.Most data on the internet platform is unstructured and textual.How to get investor sentiment from the data?How to reasonably measure investor sentiment on the network platform?What is the relationship between investor sentiment and the market?These are the concerns of investors and also are problems that this research would like to solve.When measuring investor sentiment in the network platform,we also use the text sentiment analysis method.This paper first analyzes various kinds of text sentiment analysis algorithm,then puts forward a text sentiment analysis algorithm centering on Chinese characters based on Chinese semantic.Comparing with classic algorithm,this method is more effective.This paper also constructs a combined classifier to improve performance while each method can't recognize the text sentiment well.This paper collects experiment data from Eastmoney forum.When this paper constructs investor sentiment indicators,three dimensions——post content,information about the poster and the reader are considered.Seven characteristics are used to measure the investor sentiment of the poster in the paper,including post title,the influence factor of the poster,years staying in the forum,times of view,number of up vote,forward and comment.According to the regularities of daily posting distribution,this paper also builds daily investor sentiment index and carries out empirical study.Based on this,the investor sentiment driving strategy model has been constructed and this paper makes comprehensive evaluation about the result.The research result shows that there is a positive correlation between investor sentiment and the market,and the sentiment index also has a certain role in predicting the market.The investor sentiment real-time driving strategy indicates that it is a real low-risk investment plan when analyzing the result of annualized yield,max drawdown and Sharpe ratio.
Keywords/Search Tags:investor sentiment, sentiment analysis, stock forums, event driven
PDF Full Text Request
Related items