Design an Improved Model of Software Defect Prediction Model for Web Applications

Abstract

In the software industry, web applications are crucial and are frequently updated to comply with standards or to include new capabilities. However, even while testing ensures quality, the existence of faults obstructs a smooth development. Defects are caused by a number of variables, many of which must be eliminated at great cost. Determining defects at early stage of the software development process is crucial. Therefore, it is highly desirable to develop such a model that can detect defects in a web application. In this study, the researcher surveyed object-oriented metrics and many Software Defect Prediction (SDP) models for web applications. The author has proposed an architecture for an enhanced model of SDP using machine learning techniques. The proposed model will use supervised and unsupervised learning on object oriented metrices to overcome the problem of class imbalance, cost sensitivity and correlation between Attributes and Faults.