Performance analysis of different software reliability prediction methods

Software has gained popularity in daily activities ranging from small scale applications running on handheld devices to complex application and big data processing. The software is critical in nature as it has become the most vital part of a system resulting in risks related to software failures. The risk estimate associated with a system can be calculated using different techniques. The performance of these techniques in predicting performance has not been satisfactory under different system parameters defined in advance. A very important aspect of a software system is to monitor the behaviour of the software across different platforms. Software reliability is an important domain in monitoring and managing performance of a software system. Therefore, the need of the hour is to predict software reliability comprehensively using all scientifically acquired data sets. In this paper comprehensive analysis of various parametric and non-parametric reliability growth models has been performed. The results give an insight insight into the effectiveness of non-parametric model while calculating software reliability. This paper further justifies the importance of neural network based models in calculating reliability prediction of a software system.