Relative similarity and stability in FCA pattern structures using game theory

Abstract

Pattern Structure is a generalization of Formal Concept Analysis, comprising of objects and their description forming pattern concepts. Stability is an important measure to rank such concepts. However, in relation to similarity, stability is measured by a crisp computation. In this paper, a new Game Theory based technique is proposed to compute relatively stable concepts with high similarity. It is shown that the new proposed method can also be extended to multidimensional patterns.