Software Reliability Prediction: Derivation of Model Parameters from Failure Data
Peter David Mellor · International Journal of Quality & Reliability Management · 1987
The use of “fault‐counting” models with “discrete” data in the case of commercial software has considerable advantages for the vendor. The adapted Littlewood Stochastic Reliability Growth model has the advantage of allowing a variety of fault manifestation rates. The process of inferring the parameters of this model is presented graphically in a way intended to clarify untuitively some of the problems commonly experienced with estimation, particularly where long‐term predictions are required. Based on this, alternative objective functions are suggested for fitting the model to failure data.