Integrating neural networks with software reliability
Deepak Kumar, Yogita Kansal, P. K. Kapur · International Conference on Computing for Sustainable Global Development · 2016
In today's era, where technology is growing in a fast manner and its usage has become very common in daily routine, the reliability of the software becomes primary concern of every user and developer. Thus, predicting the reliability of software before its release acts as a deciding factor to estimate whether a vendor must produce it or not. As the analytical models rely on the assumptions which are considered before project starts, making a unique assumptions each time is quite a difficult task, hence this paper, resolves the problem of assumptions. The main concern is to develop general prediction models. For achieving the goal, failure history from similar projects is collected and then is applied to construct various neural network models for prediction. The proposed methodology simply shows how different data sets can be applied to the NN models and finds the optimal among them.