Font Size: a A A

A Hybrid Stock Selection Model Based On Stock Price Prediction And Investor Sentiment And Its Application

Posted on:2021-02-07Degree:MasterType:Thesis
Country:ChinaCandidate:B Y WangFull Text:PDF
GTID:2439330611466852Subject:Management Science and Engineering
Abstract/Summary:PDF Full Text Request
Stock selection refers to the process of selecting stocks from the same or different stock categories.It is the basis of the formation of investment strategy and the development of risk management model.The core part is the stock scoring mechanism to evaluate the value of stocks,so as to select the stocks with the highest value(i.e.the stocks with high potential income and low potential risk),and the important part is the selection of factors,i.e.the selection of input variables.In reality,the future information and investor sentiment of stock will affect the change of stock returns,and then affect stock selection.Therefore,based on the existing research and from the perspective of factor selection,this paper puts forward a mixed stock selection model based on stock price prediction and investor sentiment,and studies the impact of expected earnings and investor sentiment on stock selection.The main research contents are as follows:Firstly,a fuzzy approximate support vector regression(fpsvr)model based on signal-to-noise ratio is constructed.Firstly,considering the influence of noise information contained in the stock price trend on the stock price prediction,the characteristic variable of signal-to-noise ratio is constructed and included in the input variable;secondly,based on the original SVR model,considering the influence of two different prediction errors on the stock price prediction,the fuzzy membership degree and bilateral weight measurement method are introduced to construct the fpsvr stock price prediction model,and finally,with the help of the Shanghai Shenzhen 300 component,the fpsvr stock price prediction model is constructed According to the fluctuation of stock market,the data of stock time series from 2008 to 2019 can be divided into three stages(bull market,bear market and concussion market),and three benchmark models are established for comparative analysis.The results show that: 1.Compared with the three benchmark models,the prediction error of the stock price prediction model proposed in this paper is the lowest;2.Compared with the original SVR model,the fpsvr model can better predict the stock price The stock time series in bull market and shock market stage can make accurate stock price forecast.Secondly,on the basis of stock price prediction,the stock selection model considering expected return and investor sentiment index is further constructed.First,the paper introduces the stock price prediction model in chapter three,introduces the expected return index variables,and selects turnover rate,trading volume,IPO quantity,IPO first day return rate,consumer confidence index and new A-share account number as the proxy variables of investor sentiment,selects the main components to construct investor sentiment index through the principal component analysis method,and then selects the expected return,investorsentiment index and The related financial analysis indexes are the input variables of stock selection,and considering the limitations of the original DE algorithm in facing the problem of mixed discrete continuous variables,the sigmoid method is introduced to construct the de algorithm based on sigmoid for stock selection.Finally,the empirical analysis is carried out with the stock time series data of the 300 stocks in Shanghai and Shenzhen from 2008 to 2019 as an example,and the benchmark model and statistical test are established The results show that the average rate of return(AR)and Sharpe's ratio(SR)of the model proposed in this paper are higher than all benchmark models,which is more conducive to investors' stock selection,and thus shows the effectiveness of the model.
Keywords/Search Tags:signal to noise ratio, fuzzy approximate support vector regression, stock price prediction, investor sentiment index,stock selection
PDF Full Text Request
Related items