Title: Component identification method based on semantic similarity and the cluster algorithm

Authors: Liyan Chen; Long Tan; Jun Lu

Addresses: School of Computer Science and Technology, Heilongjiang University, Harbin, China ' School of Computer Science and Technology, Heilongjiang University, Harbin, China ' School of Computer Science and Technology, Heilongjiang University, Harbin, China

Abstract: Component identification is a key problem in software reuse. In order to obtain a set of business components (BCs) with high reuse value and good reuse performance to support reuse, a BC design method based on the cluster algorithm was proposed. Through analysing existing business models, element composite models were described to divide the domain by analysing the conception semantics of the transaction field. The hierarchical clustering analysis technique based on the similarity degree among activities was also given. In the identification process, the concept of business element similarity which can overcome the limitation of the domain platform was given. Commonality, variability, granularity, and reuse cost were taken into account in the method. Experiment results show that the valuation and performance of reusability for the transaction component are improved effectively, especially the design in the platform independent model.

Keywords: reusable business components; component identification; semantic similarity; cluster algorithms; hierarchical clustering; software reuse; semantics; commonality; variability; granularity; reuse cost.

DOI: 10.1504/IJICT.2016.079122

International Journal of Information and Communication Technology, 2016 Vol.9 No.3, pp.300 - 311

Available online: 13 Sep 2016 *

Full-text access for editors Access for subscribers Purchase this article Comment on this article