A Review On Software Defects Prediction Methods

TL;DR

This paper reviews machine learning methods for software defect prediction, highlighting neural networks and gradient boosting as the most effective algorithms based on analysis of NASA datasets.

Contribution

It provides a comprehensive review of current defect prediction techniques and evaluates their performance on multiple datasets, emphasizing the effectiveness of neural networks and gradient boosting.

Findings

Neural Networks outperform other algorithms in defect prediction.

Gradient Boosting classifiers show high accuracy in defect classification.

Performance varies across different datasets and algorithms.

Abstract

Software quality is one of the essential aspects of a software. With increasing demand, software designs are becoming more complex, increasing the probability of software defects. Testers improve the quality of software by fixing defects. Hence the analysis of defects significantly improves software quality. The complexity of software also results in a higher number of defects, and thus manual detection can become a very time-consuming process. This gave researchers incentives to develop techniques for automatic software defects detection. In this paper, we try to analyze the state of the art machine learning algorithms' performance for software defect classification. We used seven datasets from the NASA promise dataset repository for this research work. The performance of Neural Networks and Gradient Boosting classifier dominated other algorithms.