Software Reliability Prediction Based on Radial Basis Function Neural Network
Pravas Ranjan Bal, Durga Prasad Mohapatra · Advances in intelligent systems and computing · 2016
Reliability of software is the key factor of software quality estimation during the testing period of software. This paper proposes a nonparametric method using radial basis function neural network for predicting software reliability. The Bayesian Regularization method is applied in the proposed model to improve the generalization and to avoid the over fitting problem. The proposed scheme has been tested on five benchmark datasets. The results of the system are compared with other states of the traditional method and it is observed that the proposed architecture outperforms its competent systems. The results of the proposed architecture have been adequately tested with a single feed-forward neural network model and a linear parametric software reliability growth model. The experimental result shows that the radial basis function of neural network yields better performance than the traditional software reliability growth model (SRGM).