ABSTRACT
A large number of execution-time based reliability growth models have been proposed for estimating reliability of software systems. One of the key assumptions made in almost all of the models is that the complete code for the system is available before testing starts and that the code remains frozen during testing. However, this assumption is often violated in large software projects because usually the code is developed in parts. This paper demonstrates the applicability of the neural network approach to the problem of developing an extended software reliability growth model in the face of continuous code churn. In this preliminary study, neural network reliability models with and without the code churn information are compared using a data set from a large telecommunication system. The results suggest that the neural network model with the code churn information is capable of providing a more accurate prediction of future faults than the model without the code churn information.
