Extraction of Fuzzy Rules from Incomplete Data with “Do Not Care” and “Lost” Values by Rough Sets

Proceedings of ‏The 11th International Conference on Research in Science and Technology

Year: 2021

DOI:

[Fulltext PDF]

Extraction of Fuzzy Rules from Incomplete Data with “Do Not Care” and “Lost” Values by Rough Sets

Gülnur YILDIZDAN, Mehmet KAYA

 

ABSTRACT: 

Rough Set Theory (RST) is a mathematical method used in reasoning and information extraction for expert systems. RST makes incomplete, inadequate or ambiguous information appropriate for data analysis by editing it. Today, incomplete data are found in many datasets. Extracting rules from these incomplete data, which are frequently found in disease data, is extremely important in the diagnosis of diseases. In this study, an algorithm previously proposed and extracting fuzzy rules from datasets containing only “do not care” missing attribute value by RST was developed in a way that it can extract fuzzy rules from datasets containing missing attribute value in both “do not care” and “lost” type. The algorithm developed was applied to the dataset of thyroid disease and certain and possible fuzzy rules were obtained for the diagnosis of the disease. The performance of the algorithm was investigated on six different datasets that had “do not care” and “lost” kinds of missing attribute values in different numbers. It was found that the algorithm generally produced successful and consistent rules in the datasets that had “do not care” and “lost” missing attribute values.

Keywords: Rough set theory, incomplete data, fuzzy rule, rule extraction, thyroid disease.