The contribution deals with the problem of feature selection strategies for the case of high dimensional feature space. In this paper an effective method of feature selection is proposed where two different measures are considered and time and storage complexity is acceptable.