Designing an improved Adaptive Neuro-Fuzzy Inference System based on Whale Optimization Algorithm to predict blood donation

Artificial neural networks and fuzzy sets theory is one of appropriate techniques to solve engineering problems in order to predict variables of supply chains and, also, of systems with high complexity and implicitly which provide no sufficient data. The main advantage of this technique over others, which lies in the short time of data examination and algorithm discovery, is in the line with that prediction and/or its influence on the future. Adaptive neuro-fuzzy inference system (ANFIS) combines neural networks with fuzzy logic concepts and is able to use capabilities of both in one framework, the inference system of which is in conformity to fuzzy "if-then" rules having potential for learning how to approximate non-linear functions. Among applications of this technique are to define variables based on the past data and their impacts on the past temporal sequences in order to predict future conditions. This research, therefore, uses neuro-fuzzy technique in order to blood donation based on data from the past years. Since each technique has its own error rates, Metaheuristic Whale Algorithm is used to reduce errors of ANFIS by improving the parameter values of neuro-fuzzy systems. The obtained results show a reduction of the RMSE of prediction from 0.00261 to 0.00153 in the ANFIS-WOA and a 41% improvement over the ANFIS method.