Klasifikasi Software Defect Prediction Dengan Menggunakan Teknik Machine Learning

Abstract

In software development, the most important thing to consider in order to improve software quality is to ensure that there are no defects. Improving quality and minimizing software lifecycle costs has become a Software Engineering (SE) practitioner and researcher. Time and costs can be saved if defects are detected at an early stage of development. One concept to deal with such problems is Software Defect Prediction (SDP). Techniques that are often used in doing SDP are Machine Learning and Deep Learning. Naïve Bayes, Random Forest, and Decision Tree are some of the popular machine learning algorithms for SDP. Defect prediction using these three algorithms uses three datasets namely PC1, KC1, and KC2 and with split datasets 70 30, 80 20, and 90 10. As a result of this study, the random forest algorithm produced the highest accuracy on the PC1 dataset with a split dataset of 90 and 10 of 92%.