Discharge Modelling using Adaptive Neuro - Fuzzy Inference System

In this paper river stage discharge models using Adaptive NeuroFuzzy Inference System (ANFIS) and Linear Multiple Regression (MLR) methods have been developed. This paper also investigates the best model to forecast river discharge. From the literature it is clear that ANN models and Fuzzy logic models are quite applicable on river stage discharge modelling. Hence this present study carried out for hybrid ANFIS models. Ten ANFIS models were developed out of which best five ANFIS models are selected. The developed models were trained, tested & validated on the data of Godavari river at Rajahmundry, Dhawalaishwaram Barrage site in Andhra Pradesh. Comparing observed data and the estimated data through developed ANFIS models, it has been proved that the developed ANFIS models predicted better results the traditional models, like MLR.