Stock Selection Analysis Base On Data Mining Technology

Forty years since the reform and opening up,China's economy has developed rapidly,and the stock market has emerged in this context.Since the establishment of China's stock market,it has experienced a difficult process and gradually matured.In recent years,due to the steady development of China's economy,people's living standards have greatly improved compared with yesterday.The savings in the home range from scratch,from less to more,and the demand for investment is growing.With the emergence of many new investment methods,people's awareness of finance and investment has gradually increased.High-risk and high-return are the characteristics of the stock market.People have been attracted to it for a long time.Nowadays,buying and selling stocks has become a way of investment for many people.People put their extra money into the stock market in order to get the desired return.However,due to the lack of professional knowledge and the asymmetry of market information,most ordinary investors' investments are blind and they want to speculate in the stock market,so the gains are not very satisfactory.Therefore,for the stock market,finding an effective investment method can not only reduce the risk of people investing,but also increase the return on investment.This is very important.In the stock market,a large amount of data is generated every minute,plus the financial data that the listed company will publish regularly.How to use the data to reduce the investment risk of investors and make investors get a better return on investment is a worthwhile research questions.The act of exploring the trend of stock price changes by using accurate and scientific analytical methods is stock investment analysis.Stock investment analysis plays an important role in the securities investment process.Appropriate investment analysis can help investors make more accurate choices,help reduce investment risks and obtain better returns.Stock investment analysis research is to maximize the utility of stock investment,how to minimize the risk under the same profit situation,and how to maximize the profit under certain risk conditions.To achieve these two points,we need to use professional analysis methods to comprehensively consider the factors affecting the value and price of stocks,so as to make objective and accurate judgments and correct decisions.Traditional statistical models have higher requirements for data and more assumptions about data.In the actual stock market,data is often difficult to meet such requirements.In contrast,data mining technology has much lower data requirements,and it can better handle non-stationary,non-normal,and high-noise data.Moreover,data mining technology can continuously acquire new data and dynamically update the model,which is very suitable for new environments.Based on this,this paper chooses the data mining method to explore the relationship between financial indicators and stock investment value through the company's financial indicator data.Whether the listed company's financial statement data disclosed to the public is helpful to the decision of ordinary investors and how to make better use of this information is the research significance and concern of this paper.This paper uses data mining related technology to empirically analyze the relationship between the financial indicators of listed companies and the investment value of listed companies' stocks.In this use,decision trees,neural networks,and logistic regression in data mining methods are obtained.The financial indicators of listed companies are the input variables of the model studied in this paper.The hidden value is set as the stock equity win rate,and the previous data is used to build the model.This paper uses three methods to build the model,and has achieved certain predictive effects.To some extent,it proves that there is an intrinsic relationship between the financial indicators of listed companies and the rise and fall of stocks.Therefore,when I invest in medium and long-term value of stocks,The company should be considered to publicly disclose financial indicators,so as to more accurately determine the rising and falling trend of stock prices.This paper uses the relevant technology of data mining to empirically analyze the relationship between the financial indicators of listed companies and the value of stock investment.Among them,decision trees,neural networks,and logistic regression models in data mining methods are used.The financial indicators of listed companies are the input variables of the research model of this paper.The target variable is set to the individual stock equity.The following conclusions are drawn:First,the relationship between the financial indicators of listed companies and the stock value does exist.Medium and long-term value investment should consider the financial indicators publicly disclosed by listed companies,so as to more accurately judge the rise and fall of stock prices.Second,among the many financial indicators of the company,some indicators have a greater impact on the company's stock value.Among them,we should focus on:net profit growth rate,earnings per share(basic)and return on equity(diluted),etc.,investors can focus on these indicators in the process of investment.Third,the paper compares the prediction effects of the three models.Among the three methods used,the neural network has the best prediction effect,followed by logistic regression and decision tree.The innovation of this paper is that the stocks are compared with the market's ups and downs as an important reference to select stocks with investment value;using data longer than previous studies,choose the past five years from 2013 to 2017.The historical data is modeled;SAS statistical software is selected as a tool for data mining,making full use of its advantages in automatic modeling and data processing.