CLUSTERING TECHNIQUES IN FINANCIAL DATA ANALYSIS APPLICATIONS ON THE U.S. FINANCIAL MARKET

In the economic and financial analysis, the need to classify companies in terms of categories, the delimitation of which has to be clear and natural occurs frequently. The differentiation of companies by categories is performed according to the economic and financial indicators which are associated to the above. The clustering algorithms are a very powerful tool in identifying the classes of companies based on the information provided by the indicators associated to them. The last decade imposed to the economic and financial practice the use of economic value added as an indicator of synthesis of the entire activity of a company. Our study uses a sample of 106 companies in four different fields of activity; each company is identified by: Economic Value Added, Net Income, Current Sales, Equity and Stock Price. Using the ascending hierarchical classification methods and the partitioning classification methods, as well as Ward’s method and kmeans algorithm, we identified on the considered sample an information structure consisting of 5 rating classes