Component Based Software Quality Prediction Using Soft Computing Techniques

Abstract

newline In today s technological world the component-based software system is applied in newlinevarious fields of engineering, to ease their multitasking function. In component-based newlinesoftware engineering, the software functions are created as components, it can reuse for newlinedifferent purpose. In modern industries, the software components took a major role to newlinecontrol the machinery. So, the quality of these software component should ensure, newlinebefore implementation. In the beginning the techniques such as ISO 9126, Bertoa, ISO newline25010, CBQM and Alvaro model are utilized to analysis the quality of software. The newlineprediction performance of these quality measures is not up to the mark with newlineconventional approaches, hence this study motivated to develop a soft computing-based newlineprediction approach for the software quality prediction. This study proposed four newlinedifferent approaches for the software quality prediction. newlineIn the first approach an improved particle swarm optimization based neural network is newlinepresented for the prediction of software quality. The IPSO proposed to tune the neural newlinenetwork based on the reliability value. Then to enhance the prediction performance the newlinesecond approach proposed an optimal fuzzy classifier based with evolutionary newlineprogramming. The evolutionary programming is used for the optimal fuzzy rule newlineselection, so that the decision accuracy of fuzzy can enhance. In comparison with the newlineprevious IPSO, the optimal fuzzy classifier provided better prediction performance such newlineas reliability and maintainability.