Wavelet Analysis

The international dry bulk shipping market is a very important component part of the international shipping market, for it has a great market share in both supply and demand of the international shipping market, and more than half of the cargoes transported in this market are natural resources that are essential to the development of a country. The freight index of dry bulk shipping market is a concrete represent for the international dry bulk shipping market, and it is a reference for those parties that are engaged in this market to make good decisions for their business activities. During recent years, the dry bulk freight index remained sharp fluctuations for it was more sensitive to incidents because of globalization. So the fluctuations of freights brought great opportunities to the participants, meanwhile accompanied with tremendous risks.First of all this paper describes recent dry bulk shipping market, and then analyzes the international dry bulk shipping market in depth from the angle of market supply and market demand, discussing the factors that would have influence on market supply and market demand at the same time. The dry bulk freight index is an outcome from the competition of market supply and market demand. So the paper reaches some conclusions of the fluctuation features by analyzing the recent dry bulk freight index data and the causes that the freight index remains sharp fluctuations. The analysis described above lays a foundation for further research.This paper uses wavelet analysis and support vector machine (SVM) combined to forecast the freight index of international dry bulk shipping market. The freight index is the outcome of a complex nonlinear system, so we cannot get sound results by using classical forecasting models. Theoretical studies have showed that SVM which based on the statistical learning theory can form good regression models for complex nonlinear systems, and avoid the defect artificial neural network has. In order to improve the forecast accuracy, this paper makes use of wavelet analysis to eliminate the impact on the market which is caused by random incidents. The outcome of the forecasting model indicates that the model this paper brings forward can reach a high accuracy prediction by grasping the character of freight index fluctuation. Meanwhile, this forecasting model is applicable to other index forecasting.