PŘÍNOS UČÍCÍCH SE METOD STATISTICKÉHO ROZPOZNÁVÁNÍ OBRAZU PŘI HLEDÁNÍ FAKTORU KONKURENCESCHOPNOSTI ČESKÝCH PODNIKŮ
Contribution of statistical pattern recognition methods in the search of competitiveness factors of Czech companies
Languages of publication
The submitted paper concentrates on the methodical aspects of measuring the relationship between potential competitiveness factors and the corporate competitiveness. The authors employ methods of statistical pattern recognition, particularly the sequential forward flow search algorithm (SFFS). The algorithm is applied on data from 432 companies. For these companies there was known their financial performance and there was up to 683 (each company) potential factors of this performance in authors ´s database. The text therefore summarizes the known approaches, describes the SFFS algorithm and proves its contribution to this field of research. An undeniable advantage of this method is its low demands on data: it does not require the normality or an a priori model. Also, it is able to evaluate relationships among many variables at once in acceptable time frame. The article presents the drawbacks of this method as well.
922 – 937
Publication order reference