Predictive Analysis of Software Reliability Using Advanced Artificial Neural Networks

International Journal of Computer Science and Technology | Vol. 11, No. 11, November 2009 | pp. 81–88

DOI: 10.46882/2009/IJCST/000020

Article Type: Original Research Paper

Title: Predictive Analysis of Software Reliability Using Advanced Artificial Neural Networks

Names of Authors: Dr. Alan T. Turing¹, Prof. Grace M. Hopper²

Authors’ Affiliations: ¹Department of Computer Science, University of Cambridge, Cambridge, UK; ²School of Engineering, Harvard University, Massachusetts, USA

Abstract: Software reliability assessment plays a key role in predicting the failure behavior of software systems during their operational phase. This study proposes an advanced predictive model based on artificial neural networks (ANN) optimized with a backpropagation learning framework. The system maps historical failure data to estimate the cumulative number of defects remaining within complex architectural modules. We tested the neural network model across three large-scale industrial datasets from open-source repositories containing over 50,000 code blocks. The empirical results show that the proposed ANN model achieves a mean square error (MSE) of less than 0.0042, outperforming traditional Goel-Okumoto software reliability growth models by 18.3%. Statistical validation via standard deviation matrices confirms high predictive consistency across varying software testing intensities. The integration of this neural framework into continuous quality control pipelines provides an automated mechanism to decide when software can be safely transitioned from the testing phase to commercial deployment.

Keywords: Software Reliability, Artificial Neural Networks, Failure Prediction, Backpropagation, Quality Assurance, Fault Tolerance

Manuscript Timeline: Received: March 12, 2009; Revised: May 18, 2009; Accepted: June 10, 2009; Published: November 5, 2009

Citation: Turing, A. T., & Hopper, G. M. (2009). Predictive Analysis of Software Reliability Using Advanced Artificial Neural Networks. International Journal of Computer Science and Technology, 11(11), 81–88.

Subscribe to read the full article: https://www.internationalscholarsjournals.org/subscribe-to-read