Estimation of parameters for S-shaped software reliability growth models according to data collected during previous releases

This article is devoted to the selection of parameters for modeling software reliability. This article discusses the process of selecting the parameters S-shaped growth model of software reliability based on data on failures in previous releases, compares the accuracy of the model at different ways of selection parameters. Goel-Okumoto model is used as the base model. This model is based on data about failures in the program's over a certain period of time. In order to obtain appropriate estimates using this model requires a certain amount of failure data, which are not available until the system has not been tested for a long enough period of time. As the experimental data using statistical data collected for three consecutive releases of the software industry scale. Evaluation of model parameters was performed by using the maximum likelihood function.