Predicting software reliability with spiking neural networks

Abstract

Software reliability is in demand as the software development is increasing rapidly. The fundamental issue of software reliability isn't having legitimate arrangement and there might be an issue with plan moreover. The impact of setup shortcomings is a result of its lack of quality, which in this manner rises up out of human research failures. In the past, many techniques are applied to predict software reliability and gaining the result may not be as effective as it goes because the data is increasing. We employed a new technique called spiking neural network to predict the software reliability. The three failure datasets are taken from literature, those are Musa 1979a, Musa 1979b, Iyer and Lee 1996. The final result is calculated in terms of Normalized Root Mean Square Error (NRSME) values.