Exploring Principal Component Analysis in Defect Prediction: A Survey

The performance of the defect prediction is solely based on the dataset which consist of software metrics. The software metrics or features sometimes very huge in number that makes the dataset complicated and it impacts the classifier or regressor efficiency. The dimension reduction technique is used to solve the problem of massive dimension of features. One of the most commonly used methods to deal with this issue is Principal Component Analysis (PCA). It is a statistical technique used for dimensionality reduction of the vast dataset in machine learning. Large number of research has been taken to predict the defective modules using principal component analysis. Its main function is to reduce the large number of features by extracting the uncorrelated features into groups. It helps to get simpler dataset, easy to handle and visualize, sometime in the cost of accuracy. In line with this, computation complexity of the predicators reduces and takes less time for execution. This survey paper is the exploration of the various studies conducted using principal component analysis for the defect prediction. The survey presented based on 28 studies from 2002 to 2020. It includes topics related to pre-release or post-release defects in software and machine equipment’s related defects. The primary focus of this study is to examine the significance of principal component analysis, its contribution in defect prediction and to identify the expansion in this topic. The study will provide helpful outlines of this subject to on-topic scholars and experts.To view and download the paper for free, visit: http://pices-journal.com/ojs/index.php/pices/article/view/24