查看文章

In this paper, we proposed a new genetic clustering algorithm based aspect mining. Aspect mining tries to identify crosscutting concerns in a software system. Clustering in aspect mining deals with grouping of methods in a software system where each group represents a crosscutting concern. Our approach stands on adjusting a standard genetic algorithm to solve clustering problems, which employed the genetic algorithm operators with suitable probability values. We used two vector space models to associate method and its attribute values. The distance between two methods is expressed using Euclidian distance. In our approach to calculate methods attribute values, we used open source library ASM [14], to analyze compiled java classes.