Common-possible concept analysis: A granule description viewpoint

Concepts are basic units of human cognitive activities and concept based granule description is one of the possible ways to realize explainable AI through information granules. However, the existing types of concepts cannot work effectively when it is necessary to simultaneously investigate the common attributes and possible attributes of granules. In order to solve this problem, we propose common-possible concept analysis based on a granule description viewpoint. Concretely, common-possible concepts are proposed, which can concurrently describe the common attributes and possible attributes of granules. Then, a method to acquire common-possible concepts from formal contexts is described. Besides, the connections between common-possible concepts and the other types of concepts are explored. Analysis results show that common-possible concept analysis can effectively deal with a type of problems and provides some details of granules different from that of formal concepts and property oriented concepts.

Automated summary

A new method of common-possible concept analysis is proposed in this paper, which can effectively describe the common attributes and possible attributes of granules, addressing the shortcomings of existing types of concepts in this aspect. Experimental results show that this method can effectively deal with a type of problems and provide granule details different from that of formal concepts and property-oriented concepts.