A Study on Learning Factor Selection and Test Resource Effect

The hyper-geometric distribution software reliability growth model (HGDM), one of the recently developed software reliability growth models (SRGMs), has been successfully applied to the estimation of the initial number of faults in a software. Learning factor plays a principal role in HGDM, which is regarded as a parameter to be estimated. So far, various plausible deterministic learning factors have been devised and proposed. However, the previous researches on HGDM do not present how to select an appropriate learning factor from candidate learning factors. We thus suggest some general learning factors and a selection procedure. A numerical example is given to illustrate the procedure. In order to reduce the number of candidate learning factors, we empirically study the additivity of test resource effect by analyzing several real test data sets.