A Study on Modeling Techniques in Software Fault Prediction

Software fault prediction is a key area in the field of software engineering. Fault can occur in any stage of software development but if it is not correctly identified and removed, it may lead to system failure. The main purpose of software fault prediction is to find faults prior to the testing phase, which will further decrease the time and cost of developing software and upgrade the quality of software. In this paper, we review machine learning and ensemble-based modeling techniques concerning software fault prediction. This paper will help the researchers to choose the best modeling technique with respect to datasets and performance measures. The findings reveal that the presence of various data quality issues related to datasets degrades the accuracy of fault prediction and also find that ensemble-based techniques are more efficient as compared to machine learning techniques.