Test-effort dependent software reliability measurement
Shigeru Yamada, Jun Hishitani, Shunji Osaki · International Journal of Systems Science · 1991
Software reliability growth models are developed here in order to relate software reliability to the effect of the test-effort spent on software testing. These models are based on non-homogeneous Poisson processes (NHPPs) which describe a software error detection phenomenon. The test-effort functions are expressed by exponential and Rayleigh curves. The estimators for the test-effort parameters are given by the method of least-squares, and those for the reliability growth parameters in the NHPP models by the method of maximum-likelihood. Using actual test data, software reliability analyses are illustrated.