关键词: 水路运输, SVR, 小样本, 损害赔偿, 预测模型

Abstract:

The bulk chemical ship is a risk of shifting dangerous source which is the greatest threat to marine safety and environment in transit when the accident occurs. Therefore, it is necessary to combine the characteristics of the chemicals in the transport for the reasonable estimate the amount of chemical spill, pollution losses for ship pollution claims and environmental protection based on the scientific foundation. SVR is a nonlinear regression model on the basis of statistical learning theory, and it has good versatility and robustness. The establishment topology is based on a small subset extracting from the training data by algorithm. The optimal solution will be obtained under the limited sample. Because of the above feature, SVR has been widely used in areas such as regression and fitting. Firstly, the basic regression and fitting principles of the SVR is introduced. Secondly, the accident damage compensation is predicted by analyzing the different examples and selecting the different kernel function. Finally, the results show that the damage compensation model is valid and feasible.

Key words: waterway transportation； SVR, small sample, damage compensation, prediction model