Automated Detection of Software Defects Through Machine Learning and Source Code Analysis

Daniel Lim Wei Jian

Automated Detection of Software Defects Through Machine Learning and Source Code Analysis

Author(s)	Daniel Lim Wei Jian
Country	Malaysia
Abstract	Software defects represent a persistent challenge in software engineering because incorrect program behavior can affect system reliability, maintainability, operational efficiency, and user confidence. Conventional software quality assurance relies on techniques such as code inspection, static analysis, and software testing to identify potential problems before deployment. Although these approaches provide essential mechanisms for evaluating software quality, large and frequently changing codebases can make comprehensive inspection resource-intensive. Machine learning introduces opportunities to analyze historical defect information and source code characteristics to identify components that may require additional verification. The present paper develops a hypothetical methodological framework for examining automated software defect detection through source code analysis and machine learning. The investigation focuses on predicting defect-prone source files rather than claiming to identify the exact location of every programming error. Four approaches are considered: rule-based static analysis, Logistic Regression, Random Forest, and Gradient Boosting. A fictional dataset containing 12,000 source files distributed across 30 software repositories is introduced to demonstrate the proposed research design. Static code complexity, coupling, code churn, and analyzer-generated warning counts are considered principal explanatory characteristics. The proposed methodology integrates source code parsing, metric extraction, historical defect labeling, feature preprocessing, project-independent validation, and comparative classification assessment. Precision, recall, and F1-score are selected as principal performance measures. Two illustrative analytical tables present the hypothetical defect-label distribution and comparative classification outcomes. All numerical values are manually constructed and do not represent actual source code analysis, trained machine learning models, or completed experiments. The methodological discussion examines defect-label reliability, class imbalance, project-specific variation, temporal information leakage, false-positive predictions, and model interpretability. Particular attention is given to distinguishing predicted defect risk from confirmed software defects. The paper concludes that combining machine learning with source code analysis provides a relevant methodological approach for prioritizing software quality assurance activities, although genuine effectiveness requires reliable defect records, reproducible experimentation, independent validation, and developer-centered assessment.
Keywords	Software Defect Detection, Machine Learning, Source Code Analysis, Software Quality Assurance, Static Code Analysis, Defect Prediction, Random Forest, Gradient Boosting, Code Complexity, Software Reliability.
Field	Engineering
Published In	Volume 2, Issue 5, September-October 2026
Published On	2026-10-02

View / Download PDF File

E-ISSN: 3067-5073

doi

CrossRef DOI is assigned to each research paper published in our journal.

IJMLAIDSE DOI prefix is
10.00000/IJMLAIDSE

Downloads

Research Paper Format Copyright Permission and Undertaking Form

All research papers published on this website are licensed under Creative Commons Attribution-ShareAlike 4.0 International License, and all rights belong to their respective authors/researchers.

CC-BY-SA

About IJMLAIDSE Fees & Payment Current Issue Publication Archive	Submit Research Paper Track Submission Status Publication Guidelines Publication Ethics Peer Review & Plagiarism	Join as a Reviewer Editors & Reviewers Get Reviewer Membership Certi.	Website/Journal Policies Usage Policy Content Policies Privacy Policy

Contact Us	+1-872-302-3851	editor@uscse.com

International Journal of Machine Learning, AI & Data Science Evolution

A Widely Indexed Open Access Peer Reviewed Multidisciplinary Monthly Scholarly International Journal

Automated Detection of Software Defects Through Machine Learning and Source Code Analysis

Share this