Evaluating extant uranium: Linguistic reasoning by fuzzy artificial neural networks

Abstract

This paper aims at estimating the extant uranium by soft computing approach. The rising contribution of this resource in the energy cycle is the reason to this research. Untidy relations and uncertain values in geological data increase the complexity of estimating extant uranium, and thus it requires a proper approach. This paper applies artificial neural networks (ANNs), in both crisp and fuzzy concepts, with the exploit of genetic algorithms (GAs). Artificial neural networks (ANNs) trace the untidy relations even though under uncertain circumstances by fuzzy artificial neural networks (FANNs), where GAs can explore the best performance of these networks. We use the type-3 of FANNs against the conventional ANNs to reveal the results, where the Lilliefors and Pearson statistical tests validate them for two geological datasets. The results showed the type-3 of FANNs is preferred for desired outcome with uncertain values, while ANNs are unable to deliver this particular.