Different Ranking of NHPP Software Reliability Growth Models with Generalised Measure and Predictability
Nguyen Hung-Cuong, Huỳnh Quyết Thắng, Le Hai-Trieu · International Journal of Applied Information Systems · 2014
In recent years, the authors have proposed a number of Nonhomogeneous Poisson process (NHPP) software reliability growth models (SRGMs) to analyse and measure the growth of software reliability during production process.This study works with a mathematics methodology to evaluate and then rank some basic NHPP SRGMs.Characteristics of ranking are: the fit of calculated occurrence failure times with real occurrence failure times; and the fit of a predicted time of the next failure with a real one.From a set of individual measures, the methodology mentioned will be deployed.The implementation of some SRGMs with several real data sets confirms that the ranking of SRGMs depends on the data sets.