A multi-granularity rough set algorithm for attribute reduction through particles particle swarm optimization

Multi-granularity analysis is an important manifestation of human cognitive ability which can obtain more reasonable and satisfactory solutions from multiple perspectives and multiple levels. In this paper, we present a multi-knowledge rapid reduction strategy combined with particle swarm algorithm based on optimistic multi-granularity rough sets. We evaluated the performance of our approach using the Statlog (Heart) Data Set and the corresponding computational experiments were completed. The empirical results indicate that our approach is effective at identifying multiple factors and usually produces multiple candidate reducts.