Hybrid Approach for the Classification of Candidate Components

Abstract

Software reuse is a way of increasing productivity and quality of a software system. An important prerequisite for successful software reuse is the choice of the most suitable components. Main problem encountered when reusing the component libraries is component retrieval; i.e. finding the components in the library that can be used in the construction of a specific information system. The formal specifications represent software that has been implemented and verified for correctness. Case-base Reasoning (CBR) system use various techniques to match a situation as a problem description, i.e. a case, to a database and known case. The goal of case retrieval is to return case that is most similar to the input attributes. Lastly rough-fuzzy hybridization mainly focuses to discover redundancies and dependencies between given features of a data to be classified The approach presented in this thesis is not fully dependent on any one concept, in stead benefits of all the three concepts together in the retrieval and classification of a candidate component. In the present work, a case-generation approach using rough-fuzzy hybridization for retrieving the candidate component used in the previous projects to be reused in the query application has been proposed. The approach is based on rough-fuzzy hybridization and structural matching and a model is proposed on the basis of this approach. The proposed model reduces the domain search and helps in retrieving the best suitable candidate component.