An Indirect Assessment Method for Environmental Damage Compensation of Ship Oil Spill Accident

A indirect method based on genetic algorithm (GA) and support vector machine(SVM) which aimed at assessing the environmental harm of ship oil spill accident was proposed. Five principal affecting factors such as oil quantity, oil specific gravity, ocean climate, environmental sensitivity and oil cleaning situations of famous cases which were accepted by International Fund for Compensation for Oil Pollution were quantified by fuzzy mathematics to be the input feature vector, while the amends amounts were output vector. In order to attain higher accuracy, the GA was adopted to optimize all the parameters globally. The results showed that average relative error was 7.9 percent which proved the method have higher accuracy. The method have practical reference value to some extent on environmental damage amendment assessment of ship oil spill accident.