An experimental analysis of similarity measures effect for identification of software component composition of services based on use-cases

Abstract

The current business environment made helpless to rethink that how software component is developed so that in future if similar feature required then it could be reused. For the reuse process the component-based software engineering was useful. In component-based development scenario the challenging task is to identify the software logical components. In the literature there are various clustering techniques with expert judgment are available to identify logical software components. In this context, all the previous methods use the similarity measure technique for finding the software cohesion. It has been observed that if any change has been made to similarity measure then it reflects changes in value of software cohesion. So, the goal of this paper is to show the effect of similarity measure for identify the logical software component of software system. For the validation and justification various feature-based similarity measures and standard parameter (Precision, Recall and Accuracy etc.) are used for software cohesion for the design of Online Broker System (case study).