Interval Symbolic Data Analysis Applications In The Analysis Of Stock

The concept of symbolic data analysis was announced by the pioneer Dr. Diday, in the 1980s. Symbolic data analysis is a new way of data analysis which was only used to analyze the symbolic data. Symbolic data is based on the traditional data, and through a "package" deal to create a higher level data which contain more information. The interval symbolic data analysis is more common among the three types of symbolic data. After more than 20 years' effort, the theory and analysis method have had a long-term development, but it's still less used in the practice. So the study focus on the application of interval symbolic data analysis in the stock analysis, to provide a new way for the investor to analyze the stock data and make the investment strategy.First the study will focus on the basic concepts of the interval symbolic data, include interval symbolic data's definition, basic algorithms, statistics and some typical application of the interval symbolic data analysis. Beside of these there will be a introduction of function of the analysis tool SODAS2.The following contents focus on the application of interval symbolic data analysis in the area of stock analysis using the actual data.At first study the application of principle component analysis of interval symbolic data. The purpose of this study is to classify a little number of principle components from many financial indicators, through analysis of the principal component data of listed companies of different sizes, discuss the relationship between the company size and investment risk and benefits, and discuss the correlation between the financial indicators. Secondly study the application of regression analysis of interval symbolic data. The purpose of this study is through the regression of the Hushen 300 index and many other stock indexes, analyze the fitting degree and relevance between the Hushen 300 index and other indexes, and find the relationship between the trend of entire A stock market and the trend of other types of stocks or financial tools. At last study the application of clustering analysis of interval symbolic data in the stock analysis. The purpose of this study is to cluster some types of stocks from many stocks by the range of price changing percentage in a period, and using the actual dynamic data of price changing to forecast the trend of different type of stocks.