Towards Profitability on the Financial Markets; A Discriminant Analysis Approach.

In this paper we tried to group in three classes the companies listed without interruption for 6 years from Bucharest Stock Exchange. We used cluster analysis, namely an iterative method of clustering, the k-means algorithm. Using data results, we have made tests for the three classes of prediction using discriminant analysis. Fisher's functions have helped us to make predictions on the affiliation of a new listed company on one of the 3 classes of risk. In this study, emphasis was placed on the liquidity of companies, but also on how efficient are used the raw materials, the basic elements in the current financial crisis. This should give us a clearer picture of companies that are ready to get over this difficult time. Key-Words: - Discriminant analysis, Cluster analysis, Pattern recognition, Stock exchange, Portfolio analysis, Classifiers.