Current classification technologies, for example decision trees work well for pattern recognition and process control. A popular and efficient method for induction of decision trees is ID3 algorithm proposed J.R.Quinlan . The main idea of this algorithm is selecting the attribute that takes the nominal values of the average mutual information and repeating this selection procedure. ID3 algorithms and its modification C4.5 make a crisp decision tree for classification. This tree consists of nodes for detecting attributes, edges for branching by values of symbols and leaves for deciding class names to be classified.