点击次数：

论文类型：期刊论文

发表时间：2012-01-01

发表刊物：International Review on Computers and Software

收录刊物：EI、Scopus

卷号：7

期号：7

页面范围：3654-3659

ISSN号：18286003

摘要：Software architecture reflects the inner structure of software systems, therefore the recovery of software architecture has great significant for software analysis, understanding, reverse engineering and reengineering, especially for legacy systems. Since hierarchical clustering is considered as a vital activity for software clustering, a clustering algorithm called Agglomerative Information Bottleneck Based Weighted Hierarchical Clustering Algorithm (ABWHC) is given to achieve software architecture recovery. The algorithm introduces more meaningful and targeted features for both procedure-oriented and object-oriented entities. Information loss is applied to calculate the similarity between entities, and label sets are generated for entities and clusters during clustering. To evaluate the effectiveness and performance of the proposed algorithm, some experiments are conducted. The results indicate the algorithm improves the accuracy and it is more flexible and targeted for software architecture recovery. ? 2012 Praise Worthy Prize S.r.l.