Traffic and Transportation Studies · 2008 · 13 citations · 0 references
ComputingData ScienceForecastingMachine LearningTime Series AnalysisKernel MethodsSupport Vector MachinesTime Series ClassificationTime Series ForecastingData ExtrapolationFreight Rate ModelingMachine Learning-based ForecastingServiceabilityServiceability AnalysisSupport HyperplanesSupport Vector ClassificationSupport Vector Machine Classification
A forewarning model of freight rate was established based on support vector machine (SVM) and its algorithm was developed in order to analyze and forecast the alarm of freight rate fluctuation in shipping market. In accordance with shipping experts' knowledge and experiences, the actual alarm degrees of freight rate within the sample interval of shipping market were determined by CCBFI, CCFI, and BFI, which were selected as alarm criteria. The classification hyperplanes and forecasting intervals of freight rate alarm degree were obtained from SVM learning and MATLAB programming, based on sample data trained. And then the interpolation and extrapolation tests were carried on. The results of tests show that this methodology has a very good serviceability regarding to forewarning of freight rate in shipping market.