Rough-Set-Based Interrelationship Mining for Incomplete Decision Tables

Abstract

Rough-set-based interrelationship mining enables to extract characteristics by comparing the values of the same object between different attributes. To apply this interrelationship mining to incomplete decision tables with null values, in this study, we discuss the treatment of null values in interrelationships between attributes. We introduce three types of null values for interrelated condition attributes and formulate a similarity relation by such attributes with these null values.