Questionnaire analysis by decision algorithm generation

With the growing importance of marketing in Poland, questionnaire studies of companies are becoming more common. The results of such studies can be presented in table form, where firms are characterized by a set of attributes and by classifications made by experts. An important practical problem is to develop a representation of this system which is free from redundant information. The rough sets theory created by Z. Pawlak is a useful tool for such an analysis. This paper explains how to exploit rough sets methodology to represent the relationships between attributes and classes in the form of decision rules. The computer program which generates decision rules is presented and is illustrated with data from the contractors' questionnaire studies.<<ETX>>