Software Reliability Growth Testing Effort Function Model Dependent on Machine Learning and Neural Network Algorithm
Refath Farooq, Mohammad Ahsan Chishti · 2024
The quantification of dependency stands as a crucial element in programming quality. The Software Reliability Growth Model (SRGM) is a commonly utilized tool for evaluating reliability across diverse and challenging scenarios. However, the accuracy of traditional time-sensitive SRGMs may be compromised when test execution extends over a prolonged period. In response to this limitation, the SRGM approach is adapted to center around testing practices rather than relying exclusively on the progression of time. Earlier methods proposed restricted testing capabilities, but the notion of unlimited test time introduces an infinite testing scenario, rendering it impractical. To overcome this, the method introduces Neural Heterogeneous Poisson Process (NHPP) models, which operate without being bound by time limitations. Furthermore, the approach incorporates Artificial Neural Networks (ANNs) to model program failures based on available information. The application of an AI model selection method ensures precise representation of both historical and future load fitting location information Artificial Neural Networks (ANNs) showcase an adept capacity for effective generalization from the input provided to a comparable model, proficiently capturing patterns from prior instances of failure data. Through the integration of common-sense information into programming failure indicators, the suggested Testing Effort Function (TEF) and Software Reliability Growth Model (SRGM) proficiently depict extensive failure data. The incorporation of Machine Learning (ML) and Artificial Neural Networks (ANN) notably enhances the precision of traditional boundary estimation when compared to release time, thereby ensuring robust software reliability.