The Comparative Studies On Forecasting Methods For Maritime Accidents

The thesis is aimed at deepening studies on marine traffic safety, at improving the management level of the marine traffic safety, and at realizing objective management of the accidents. Fist of all, maritime accidents is general introduced, the reasons of the accidents are analyzed at the point of persons, ships, environment and management, and the characteristics of maritime accidents are explained. On this basis, forecasting models which are suitable for the national maritime accidents are expounded particularly, for example, regression analysis, grey-system and exponential smoothing method. What's more, the paper tries to use residual modification model and Markov model to ameliorate the results from the grey-system model. At last, two examples of different types given are used to validate these models and methods. On one hand, the five models' characteristics and applying conditions are compared furthermore by the results;on the other hand, it shows that the two ameliorated models make the forecasting result much more precise.