Forecasting period charter rates of VLCC tankers through neural networks: A comparison of alternative approaches

Volume-wise, seaborne crude oil represents close to 45 per cent of all internationally traded crude oil – thus remaining as the modern world primary source of energy. The usual focus in seaborne freight rate forecasting literature is the spot rate, whereas, on the other hand, a limited amount of literature has been directed towards period charter rates. To the same extent, there is a scarce amount of literature available dealing with the use of artificial neural networks (NNs) in forecasting seaborne transport market rates. This article focuses on applying NNs to period charter rates forecasting of very large crude carriers. The performance achieved for 1- and 3-year period charter rate time series by two different NN models (multi-layer perceptron and radial basis function (RBF)) is benchmarked against a more elementary performance delivered by an autoregressive integrated moving average (ARIMA) model. We find that NN modelling delivers encouraging end results outperforming the benchmark model (ARIMA). We can also point out that NN using RBFs delivers the best overall predictive performance.