Heterogeneous Mixture Model for Software Reliability Prediction

The mixed distribution model can be used for the performance prediction of software reliability indices. The purpose of the present study is to develop a heterogeneous mixture model of distributions to establish the software reliability indices that are computationally tractable. The parameter estimation is done using Expectation–Maximization (EM) algorithm and implemented on the real-time software failure observation. We examine the suggested mixture models to opt the best model by using the Goodness of Fit (GOF) tests. To determine the best model, some statistical tests such as AIC, BIC, HQC, etc. are performed. Adaptive Neuro-Fuzzy Inference System (ANFIS), which is a combination of artificial neural network of ANN and fuzzy inference system, is used for comparing the numerical results obtained for the software indices via mixture model.