The subject matter of this monograph is connected with the following two problems of data mining and knowledge discovery: Representation of knowledge, contained in a decision table, in a form which is convenient for understanding. The length of knowledge description is crucial in this case. Prediction of the value of decision attribute for a new object. The accuracy of prediction is the most important aspect of this problem. These two aims (short description and high accuracy) seem to be incompatible. However, it is known that classifiers with shorter description are often more precise. This monograph is one more confirmation of this fact.