Baltic dry index forecasting using a neuro-fuzzy inference system

Abstract We employ a Fuzzy Inference System, with a specific focus on utilizing a hybrid intelligent system known as ANFIS (Adaptive Neuro Fuzzy Inference System) to forecast the Baltic Dry Index. This system integrates the adaptive learning features of neural networks with the logical reasoning of fuzzy logic, thereby offering superior forecasting accuracy compared to single-method approaches. Our findings demonstrate the superior performance of the ANFIS model in comparison to a feed-forward neural network and two traditional models, namely AR (Autoregressive) and ARMA (Autoregressive Moving Average), in terms of Root Mean Squared Error (RMSE).