To Enhance A-KNN Clustering Algorithm for Improving Software Architecture

Abstract: Software Architecture is important factor for the development of complex and big software system. Software Architecture Decomposition is an important part in software design. Software clustering is used to cluster functions of similar type in one cluster and other are in other cluster. K-mean is the base of the clustering but it has some limitations. Many clustering methods are used for decomposition the software architecture. A-KNN cluster method is more efficient than others methods but some functions are highly coupled then cluster technique does not find out correct distance. So that need to enhancement in Euclidian distance formula based on normalization. In this paper, an enhancement has proposed in the Euclidean distance formula which has increased the cluster quality. When the cluster quality will be increase then cluster the highly coupled function properly and improve the software architecture and A-KNN will be generate the best results than previous method