A Study on Software Defect Prediction using Feature Extraction Techniques

Identification and elimination of defects in software is time and resource-consuming activity. The maintenance of a defective software system is burdensome. Software defect prediction (SDP) at an early stage of the Software Development Life Cycle (SDLC) results in quality software and reduces its development cost. In this study, a comparison is performed on nine open-source softwaresystems written in Java from PROMISE Repository using four mostly used feature extraction techniques such as Principal Component Analysis (PCA), Linear Discriminant Analysis (LDA), Kernel-based Principal Component Analysis (K-PCA) and Autoencoders with Support Vector Machine (SVM) as base machine learning classifier. The model validation is performed using a ten-fold cross-validation method and the efficiency of the model is evaluated using accuracy and ROCAUC. The results of this study indicate that Autoencoders is an effective method to reduce the dimensions of a software defect dataset successfully.