Analysis And Forecast Of Stock Market Based On Data Mining Technology

As a vital part of the securities industry,the stock market has attracted wide attention from investors.It is of great practical significance and theoretical value to seek effective stock investment method to analyze and forecast the stock market and reduce the investment risk of investors.However,because the stock market is influenced by the stock intrinsic value,market factors,political factors and macroeconomic performance and many other factors,there are no determined rules between many factors,also the stock market will produce a lot of data every day,which bring some difficulty to the stock market research.Data mining,combining with database,statistics and artificial intelligence and other disciplines,it can unearth hidden valuable information from a large number of raw data.The characteristics of the stock market determine the application of data mining method to analyze and forecast the stock market,with strong practical value and practical significance.In this paper,by collecting and sorting the stock index and the financial data of the listed companies,a series of improved algorithms is put forward which is for different problems by using the classification and clustering methods in data mining,and apply it to analyze the stock market.The main contents are as follows:(1)As for the uncertainties and lack of objectivity when dealing with data in fuzzy time series model,the fuzzy time series model based on density peak is proposed and applied to forecast the stock price.(2)Because the K-means algorithm is sensitive to the initial clustering center and is easy to fall into the local extremum,a K-means algorithm based on the kernel function artificial fish swarm algorithm is proposed and applied to the clustering analysis of the stock market.(3)Aiming at the characteristics of stock data's high dimension and complexity,a stock classification model based on factor analysis and OPTICS-Plus algorithm is proposed,which achieving the classification of the stock market.The new algorithm can effectively eliminate the redundancy of the data and improve the performance and convergence speed of the clustering.(4)In view of the high dimensionality and redundancy of the listed companies' financial data,the financial early warning model of listed companies based on Lassomethod and Logistic regression is proposed,this new method judges whether the financial condition of the listed company is in crisis and achieve the effect of early warning.This paper uses the classification algorithm and clustering algorithm in data mining to predict the price of Chinese stock market,classify the stock market,and predict the financial status of listed companies.The simulation results show that the proposed method can analyze and forecast Chinese stock market more effectively,and help investors make reasonable decisions,reduce investment risk.