Fortified Offspring Fuzzy Neural Networks Algorithm

Our research here suggests a fortified Offspring fuzzy neural networks (FOFNN) classifier developed with the aid of Fuzzy C-Means (FCM). The objective of this study concerns the selection of preprocessing techniques for the dimensionality reduction of input space. Principal component analysis (PCA) algorithm presents a pre-processing phase to the network to shape the low-dimensional input variables. Subsequently, the effectual step to handle uncertain information by type-2 fuzzy sets using Fuzzy C-Means (FCM) clustering. The proposition (condition) phase of the rules is formed by two FCM clustering algorithms, which are appealed by spending distinct values of the fuzzification coefficient successively resulting in valued type-2 membership functions. The simultaneous parametric optimization of the network by the evolutionary algorithm is finalized. The suggested classifier is applied to some machine learning datasets, and the results are compared with those provided by other classifiers reported in the literature.