Forecasting Dry Bulk Freight Index With Support Vector Machine

Because the dry bulk shipping market is affected by many uncertain factors, the freight index of the market always fluctuates sharply, and the traditional forecasting methods are not suitable to its predication. This gives difficulty for the market participants to make decisions. Thus, developing models to forecast the freight index through studying the internal laws and external influence factors, can provide a powerful tool for operators and investors to understand the market trend and avoid the price risk.This paper conducts the research on freight index forecasting for Panamax bulk carrier with Support Vector Machine (SVM). SVM fully considered the randomness of freight index and its generalization ability has significantly improved compared with the neural network model, therefore, the SVM-based model is more suitable for nonlinear time series forecasting.This paper develops a model to forecast the freight index through studying the internal mechanism and the external influence factors. The model can provide powerful tool for the operators and investors to understand the market trend and avoid the price risk. By taking the freight index of Panamax bulk carriers as subject, firstly, in order to eliminate the impact of random incidents in dry bulk market, wavelet transform is adopted to de-noise the Baltic Panamax Index (BPI). Then, the wavelet transform and Support Vector Machine (SVM) combined model to predict BPI is established. The inputs of the model are values of the five prior consecutive monthly BPI, and the output is the sixth monthly BPI. The trained model and the forecasted results are obtained through SVM training. Finally, the numerical analysis shows that the wavelet transform and SVM combined model has higher accuracy and can be used to predict the trend of the freight rates of the Panamax bulk carriers.