Enhancing Software Reliability during Maintenance: A Comparative Study Using Artificial Neural Networks and Statistical Methods

Abstract

Software reliability is a critical factor in ensuring the consistent performance of software systems during maintenance phases. Traditional statistical methods have been widely used to predict and enhance software reliability; however, the advent of Artificial Neural Networks (ANN) offers a promising alternative due to their ability to model complex, non-linear relationships in data. This study provides a comparative analysis of ANN and statistical methods in optimizing software reliability during maintenance. We evaluate the effectiveness of both approaches using real-world maintenance data from diverse software projects. The findings reveal that while statistical methods offer robust baseline predictions, ANNs significantly outperform in scenarios involving complex dependencies and non-linearities. This research underscores the potential of integrating ANNs into software maintenance practices to enhance reliability, offering a pathway to more resilient and efficient software systems.