Object Oriented-Based Metrics to Predict Fault Proneness in Software Design

2025-04-11

AI Summary

This study addresses a critical challenge in object-oriented (OO) software fault-proneness prediction: the principled selection of the most discriminative core design metrics. To this end, we propose an empirically grounded OO metric selection framework that systematically evaluates the predictive power of multi-dimensional metrics—including coupling, cohesion, and complexity—through integrated statistical analysis, correlation testing, and regression modeling, all driven by real-world defect labels. Our key contribution is the first formal quantification of metric robustness, identifying high-performing metric subsets that exhibit both strong predictive accuracy and cross-project stability—thereby mitigating longstanding inconsistencies in metric selection across prior studies. Experimental results confirm the statistically significant predictive capability of several OO metrics for fault proneness. The findings provide a reusable, evidence-based foundation for constructing defect prediction models and guiding software quality improvement practices.

Technology Category

• Data Mining & Knowledge Management: Anomaly/Outlier Detection
• Machine Learning: Dimensionality Reduction/Feature Selection
• Knowledge Representation and Reasoning: Diagnosis and Abductive Reasoning

Application Category

• User Modeling, Personalization and Recommendation: Metrics for user behavior and evaluating success
• Web Mining and Content Analysis: Robustness and generalizability of Web mining methods
• Security and Privacy: Large-scale security measurements

Abstract

In object-oriented software design, various metrics predict software systems' fault proneness. Fault predictions can considerably improve the quality of the development process and the software product. In this paper, we look at the relationship between object-oriented software metrics and their implications on fault proneness. Such relationships can help determine metrics that help determine software faults. Studies indicate that object-oriented metrics are indeed a good predictor of software fault proneness, however, there are some differences among existing work as to which metric is most apt for predicting software faults.

Problem

• Predicting software fault proneness using object-oriented metrics
• Exploring relationships between metrics and fault prediction accuracy
• Identifying most effective metrics for fault-prone software detection

Innovation

• Object-oriented metrics predict fault proneness
• Relationships between metrics determine software faults
• Metrics improve software quality and development