Software Defect Prediction using Machine Learning Algorithms: Current State of the Art

Abstract

One of the essential exploratory fields in the software quality field is software defect prediction. Software engineering involves many ways to predict software quality assurance topics such as test effort prediction, cost prevention, error prediction, reusability prediction, prediction for safety, and consistency prediction. Though most of these predictive methods remain in the initial stage, and more study is in the forecast, many academicians and industry people have begun to work on new projects in this field.  Mechanisms to increase the efficiency of the assurance activities and allocate resources more effectively are becoming more efficient with Software Defect Prediction (SDP). In this article, the state of the art in software defects with Machine Learning algorithms is discussed.

Keywords- Software Defect Prediction, Machine Learning Algorithms, Static Metrics, Dynamic Metrics, Object-Oriented Metrics, SVM, Random Forest, Decision Tree