The process of Web service discovery identifies the most relevant services to requesters' service queries. We propose a new measure of semantic similarity integrating multiple conceptual relationships (SIMCR) for Web service discovery. The new measure enables more accurate service-request comparison by treating different conceptual relationships in ontologies such as is-a, has-a and antonomy differently. Each service or request is represented by vectors of terms (or words) that characterize both the interface signature and textual description. The overall semantic similarity is computed as a weighted aggregation of interface similarity and description similarity. The experimental results confirm the effectiveness of the proposed semantic similarity measure. As demonstrated in this study, the semantic Web service discovery method based on the proposed similarity measure outperforms existing state-of-the-art discovery methods in terms of precision, recall and Fmeasure. The proposed semantic similarity measure has wider applications such as to improve document classification or clustering, and to more accurately represent and apply knowledge in expert and intelligent systems.

Web service discovery is to find the most relevant services to satisfy the requester queries by means of similarity matching between Web services and a service requirement. We propose a novel Web service discovery method based on semantic similarity measure combing I/O similarity and context similarity. A new similarity measurement between terms is put forward as the basis of Web service similarity. Similarity between terms is represented by the similarity of two concepts containing them, and is calculated by means of the length of the shortest path between two concepts. Three kinds of linked relations are considered, i.e. ISA, HASA and antonymy relations. The experiments demonstrate that the proposed approach can improve the accuracy of concept similarity measure. Service discovery based on the similarity measure has better query performance than previous discovery methods in the precision, recall and F-measure.

Computer technology changes with each passing day. As a member of software system, the architecture of legacy system becomes complex and difficult to adapt to complex software functional requirements in the process of evolution. It has not only affected the late legacy system maintenance work, but also to the enterprise has brought serious losses. In order to solve the above problems, this paper presents a Fuzzy Hierarchical Clustering algorithm Based on Information Loss, FHCBIL. There are two improvements in FHCBIL algorithm: On the one hand, this algorithm extends the weights allocation of entity features, which solves the problem that the two element relation can not reflect the influence degree of entity feature on entity. On the other hand, FHCBIL algorithm uses the information loss as a similarity calculation method, avoiding the problem of sensitive data set to the irregular data set in the clustering process. In this paper, we further apply FHCBIL algorithm in software re-architecting, give the model of software re-architecting, and implement new nested software architecture. In order to evaluate the quality of the new architecture, we evaluate the new architecture from two aspects: the rationality of the new architecture hierarchy and the accuracy of entity division. The experimental results show that the new architecture based on FHCBIL algorithm has high cohesion and low coupling. And the accuracy of the new architecture is high.