An Adequate Representation of Medical Data Based on Partial Set Approximation

Abstract

Abstract. Computer aided medical diagnosis and treatment require an adequate representation of uncertain or imperfect medical data. There are many approaches dealing with such type of data. Pawlak proposed a new method called rough set theory. In this paper, beyond classical and recent methods, the authors propose a basically new approach. It relies on a generalization of rough set theory, namely, the partial covering of the universe of objects. It adequately reflects the partial nature of real-life problems. This new approach called the partial approximation of sets is presented as well as its medical informatics application is demonstrated.

Ranking Rough Sets in Pawlak Approximation Spaces

“…These examples deal with studying the symptoms of thyroid dysfunctions. Although the problem emerged in Csajbók et al [13], a substantially different solution is presented here.…”

“…Let us consider a data table given in Table I, taken from [13]. It contains clinical symptoms which may indicate that someone, a patient, develops hypothyroidism or hyperthyroidism, perhaps neither of them.…”