Implementing a PSO-ANFIS model for prediction of viscosity of mixed oils

ABSTRACTThis work presents a reliable model named adaptive neuro-fuzzy inference system (ANFIS) optimized by coupled simulated annealing (PSO-ANFIS) for the prediction of viscosity of mixed oils. The model was developed based on various three and two oil sample mixture data gathered from literature covering wide range of influencing parameters. Different graphical and statistical approaches were used to evaluate the model performance in estimations of actual data. Predictions of the developed model were also compared with four well-known literature models for prediction of viscosity of mixed oils. Results show that the developed particle swarm optimization ANFIS model exhibits accurate predictions and presents reliable results compared to other literature models. The model exhibits an overall AARD% value of 2.59%.