Predictive Analytics for Software Reliability: Integrating Software Metrics with Automated Model Search for Quality Assurance Optimization

Rashid Anwar Syed; Arvind Kumar Pandey

doi:10.36948/ijfmr.2026.v08i05.88953

Predictive Analytics for Software Reliability: Integrating Software Metrics with Automated Model Search for Quality Assurance Optimization

Author(s)	Mr. Rashid Anwar Syed, Dr. Arvind Kumar Pandey
Country	India
Abstract	Predicting software reliability is important but difficult: traditional statistical models and manually tuned ML approaches fail to express nonlinearities present in modern software and often demand extensive human domain knowledge. We evaluate an end-to-end automated hyperparameter search pipeline of RF, GB, and SVM classifiers soft-voted together on 21 static code metrics using the NASA KC1 dataset of defect proneness (1162 modules, 25.3% defective) compared with standalone SVM, RF, and ANN baselines under 10-fold stratified cross-validation. SMOTE oversampling is used in each training fold to account for the class-imbalanced data, as explored by recording low recall on a preliminary run with unbalanced training data. Training with balanced data, the automated ensemble meta-model outperforms the other three in terms of F1-score (0.482) and RMSE (0.411), and ties RF for largest AUC (0.722), a slight but consistent improvement upon manually tuned baselines. Halstead complexity and volume metrics, along with overall lines of code are the best predictors of defect proneness. Automated model selection appears to offer a tangible benefit in software reliability prediction, but is only a marginal improvement rather than a silver bullet. Additionally, handling class imbalance is at least as important to generating useful, balanced predictions.
Keywords	Software Reliability, Defect Prediction, Automated Machine Learning, Software Metrics, Class Imbalance, SMOTE
Field	Computer Applications
Published In	Volume 8, Issue 5, September-October 2026
Published On	2026-10-03
DOI	https://doi.org/10.36948/ijfmr.2026.v08i05.88953

View / Download PDF File

About IJFMR 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 Reviewer Referral Program Get Reviewer Membership Certi.	Website/Journal Policies Usage Policy Content Policies Privacy Policy

Contact Us		+91-9687-828-838	editor@ijfmr.com

International Journal For Multidisciplinary Research

A Widely Indexed Open Access Peer Reviewed Multidisciplinary Bi-monthly Scholarly International Journal

Predictive Analytics for Software Reliability: Integrating Software Metrics with Automated Model Search for Quality Assurance Optimization

Share this