ABSTRACT
For software engineering, software reliability prediction is crucial. The most crucial element is software reliability assessment, which is used to characterize any software products quality quantitatively throughout the testing phase. Over the years, numerous analytical models have been put forth for evaluating the dependability of software systems and predicting the development patterns of software reliability with various levels of prediction capability at various testing phases. However, it is necessary to create a single model that can be used to make predictions that are more accurate in all circumstances. To address this, the neural network (NN) method is presented. This study explains how neural-network-based models may be used to predict software dependability more accurately in the actual world and offers a method for quantifying the improvement in software reliability using neural network models. A variety of datasets with software failures are used to test the neural network models. Several software initiatives have provided these datasets. The findings demonstrate that performance is greatly enhanced when using neural network models as opposed to conventional statistical models built on a non-homogeneous Poisson process.
