<![CDATA[Defects that may arise in software during the development process can affect the quality of the software. The classification method is used to predict software defects to minimize defects. However, the dataset used in the classification process may contain less relevant or have too many features. This can be overcome by selecting features in the dataset. In this research, the Random Forest algorithm is applied for the classification process, and the Artificial Bee Colony (ABC) algorithm is used as a feature selection method. The research aims to determine the accuracy of Random Forest with ABC feature selection. From the results of research conducted on 3 Relink datasets, without feature selection, an average accuracy of 73% was obtained. After implementing ABC feature selection, the average accuracy increased to 82%.]]>
Copyrights © 2024
