Analysis and forecast of the capesize bulk carriers shipping market using Artificial Neural Networks

Abstract

Investing in the bulk carrier market constitutes a rather risky investment due to the volatility of the bulk carrier freight rates. In this study it is attempted to uncover the benefits of using Artificial Neural Networks (ANNs) in forecasting the Capesize Ore Voyage Rates from Tubarao to Rotterdam with a 145,000 dwt Bulk carrier. Initially, market analysis allows the assessment of the relation of some parameters of the dry bulk market with the evolution of freight rates. Subsequently, ANNs with an appropriate architecture are constructed and sufficient data, in terms of quantity and quality, are collected and organized so as to establish both the training and the testing data sets. The use of ANNs along with genetic algorithms allows the prediction of bulk freight rates with considerable accuracy for as long as eighteen months ahead and this is quantified by calculating the relative and absolute errors. It is concluded that ANNs offer a promising approach to forecasting the bulk market when coupled with efficient market modeling.