Research of Software Reliability Combination Model Based on Neural Net

In order to improve the accuracy of software reliability prediction, a neural network software reliability combination model is proposed from analyzing the classical software reliability models and neural network model. The proposed model combines the software reliability model and neural networks by constructing the transfer function corresponding to the selected base model. The neural network combination model can obtain a satisfactory generalization with a smaller network. The results of experiments show that the neural network combination model can effectively improve the forecasting accuracy of software reliability.