Extraction of fuzzy rules from incomplete data with do not care and lost value by rough sets

Rough set theory is a mathematical approach to imprecision, vagueness and uncertainty in data analysis. This study deals with the problem of producing a set of certain and possible fuzzy rules from incomplete quantitative data. In this thesis, knowledge extraction has been done from incomplete data sets using rough set theory. In the present, incomplete data mainly exist in medical data sets. It is extremely important to extract rules from these data for medical diagnosis. Through this thesis, a data set including two different incomplete data which are called as ?do not care? and ?lost? data was used. A novel algorithm has been proposed to extract fuzzy rules by rough sets from incomplete data with quantitative values, as apart from the previous studies which handle the only ?do not care? data or the only missing values. For this purpose, a software has been developed for evaluating the proposed algorithm on thyroid data set. The algorithm has been tested for six different cases which use various type and number of incomplete data. The experimental results demonstrates that the proposed algorithm provides the accuracy rate of 100% in certain rules and 91% in possible rules. The lowest accuracy rate has been obtained for the case in which all of incomplete values are lost. In such a case, as the number of rules increases, the accuracy rate decreases. Keywords: Rough set theory, incomplete data set, lost value, do not care value, fuzzy rules