A New Case-Based Classifier System Using Rough Formal Concept Analysis

Abstract

Rough set theory and formal concept analysis were invented by Pawlak and Wille in the 1980s and have been applied successfully in several domains. In this paper, we propose a new case-based classifier system based on an integrated rough set theory and formal concept analysis technique. We focus on the construction of a better knowledge base to produce the classification rules. Our system employs rough set theory to discover reduced cases. We then formulate a knowledge base with hierarchical structure by using formal concept analysis. The result is a concept lattice knowledge base embedded in our case-based classifier. We can generate classification rules from implications and subconcept-superconcept relations inside the obtained concept lattice. An illustrative example and a case study are provided to demonstrate the feasibility and applicability of our system. The advantages of our system are thus a better knowledge base for new problem classification and the flexibility to learn new rules.