Estimate the performance of applying machine learning algorithms to predict defects in software using weka

Predicting faults in any software becomes one of the critical study fields in software engineering. Software defect prediction in the early stages of the software development life cycle helps in improving software quality and reliability. Machine learning procedures are considered to be the most crucial technique for predicting defects. This paper measures and compares the performance of using five supervised machine learning algorithms to predict defects in software using the Weka tool. The datasets were used from the PROMISE repository. The results proved that supervised machine learning performed well in software fault prediction with an accuracy of more than 81%. Moreover, compared to the algorithm used, random forest and artificial neural networks have the highest accuracy for predicting defects.