Predicting Software Faults Based on K-Nearest Neighbors Classification

Abstract

Software defect prediction is one of the most important task during the development of software systems in order to save developers’ time and effort. Discovering defects in an early stage of software development will allow programmers and developers to take action and resolve these faulty parts in software before its launch. In this paper, the K- Nearest Neighbor (KNN) machine learning algorithm is used to predict faulty software projects. Experimental studies are conducted on five public datasets with different similarly measures. Results showed that KNN can be used to predict software faults with accuracy rate that can achieve up to 87.2%.