EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

AI-Enhanced Software Reliability Growth Model Using Machine Learning for Predictive Fault Detection and Reliability Forecasting

Mohammad Altaf Dar, Mohd Tahir Irfan, Amir Hussain

International Journal of Business & Applied Sciences · 17.09.2026 · 11 (1) · 1-9

DOI: 10.68870/ijbas.v11i1.7 ↗

Kurz gefasst

Fachartikel in der Zeitschrift International Journal of Business & Applied Sciences, online erschienen am 17.09.2026. Verfasst von Mohammad Altaf Dar und 2 weiteren Autor:innen. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Purpose

This study develops an AI-enhanced Software Reliability Growth Model (SRGM) to address a limitation of traditional SRGMs, which assume a constant fault detection rate and therefore do not adequately capture the dynamic behavior of contemporary agile and DevOps environments.

Methods

Random Forest and Long Short-Term Memory (LSTM) networks were integrated into the Non-Homogeneous Poisson Process (NHPP) framework, representing the fault detection rate, b(t), as a time-varying function of software metrics such as code churn, cyclomatic complexity, test coverage, and historical faults. The models were trained and validated on four NASA PROMISE datasets (CM1, KC1, JM1, and PC1) and a synthetic NHPP dataset. To ensure reproducibility, hyperparameters were optimized using 5-fold cross-validation, random seeds were fixed, and software and hardware details were documented.

Findings

The LSTM-based SRGM outperformed the Goel-Okumoto model (MSE 210.4, R² 0.81) and the Musa-Okumoto model (MSE 198.2, R² 0.83), achieving an MSE of 88.7 and an R² of 0.95. The final fault count prediction error was reduced from 11.2% to 2.8%, and reliable forecasting was possible after only 30% of the test history.

Implications

The framework integrates with real-time CI/CD dashboards and helps prioritize refactoring efforts, providing a practical, data-driven alternative to static SRGMs. Originality: The study embeds a machine learning-based, time-varying fault detection function directly into the NHPP differential equation, preserving the theoretical foundation of SRGMs while enabling data-driven adaptability.

Zitieren

APA

Dar, M. A., Irfan, M. T., & Hussain, A. (2026). AI-Enhanced Software Reliability Growth Model Using Machine Learning for Predictive Fault Detection and Reliability Forecasting. International Journal of Business & Applied Sciences, 11(1), 1-9. https://doi.org/10.68870/ijbas.v11i1.7

Vancouver

Dar MA, Irfan MT, Hussain A. AI-Enhanced Software Reliability Growth Model Using Machine Learning for Predictive Fault Detection and Reliability Forecasting. International Journal of Business & Applied Sciences. 2026;11(1):1-9. doi:10.68870/ijbas.v11i1.7

BibTeX
@article{Dar2026,
  author = {Dar, Mohammad Altaf and Irfan, Mohd Tahir and Hussain, Amir},
  title = {AI-Enhanced Software Reliability Growth Model Using Machine Learning for Predictive Fault Detection and Reliability Forecasting},
  journal = {International Journal of Business & Applied Sciences},
  year = {2026},
  volume = {11},
  number = {1},
  pages = {1--9},
  doi = {10.68870/ijbas.v11i1.7}
}
Kostenlos starten →