Enhancing Reusability and Measuring Performance Merits of Software Component using Data Mining

Abstract

In software industry, it is necessary to reduce time and efforts in software development. Software reusability is an important measure to improve development and quality of software. Enhancing reusability will reduce delivery time of software products, decreases the development labor, software defects and cost of the development process. Software reuse is the best solution factor to acquire the existing knowledge from software warehouse. Measuring the reusability level of the software is essential to achieve the goals of reuse. Data mining is the process of extracting useful patterns and analyzing data sets from large data sets. The reusability of a software component chooses the right measurement and enhances the preventability of an application for reuse. The software metrics are used as quantitative measures to establish and evaluate the components. In this paper, measuring the software reusability using some classification algorithms on a specific software reuse data set is discussed. The system is implemented using R data mining tool and performance of the automation system is developed for reusability prediction like precision, recall, f-measure. The experimental result shows that the model can be effectively used for efficient, accurate, quicker and economic identification of reusable components from the existing software resources. This paper seeks to provide comparative analysis of H-SOM and Naïve Bayes algorithm classifiers of Dengue datasets