| The potential risk of water traffic accident increases with the continuous and steady development of waterway transport as well as spiking number of ships at sea.Thus,it is of great practical significance to construct an appropriate prediction model to clarify the complicated relation between the number of water traffic accidents and the influencing factor,which could be beneficial to control the trend of water traffic accidents.The current research status of water traffic accident prediction methods was first summarized from three aspects:qualitative prediction,quantitative prediction and combined prediction;the theory and prediction performance evaluation method of five single prediction models and combined prediction models were introduced in good detail;Bird theory was applied to study the mechanism of water traffic accident;gray relative analysis method was used for quantitative analysis of the influencing factors of water traffic accident from human factors,ship factors,environmental factors,management factors and economic factors to identify the key influencing factor;on the basis of the comprehensive analysis characteristics of nonlinearity,randomness,dynamics and uncertainty of the water transport system and the applicability of each prediction model,gray model and BP neural network model were selected to establish a combined optimization model by error correction technique.Finally,the data of water traffic accidents during 2001～2014 were used as sample data and the data of water traffic accidents during 2015～2016 as test data;two single model and combined optimization model were used respectively for case prediction;the prediction results of the combined optimization model,as well as the gray model and the BP neural network model were compared by the prediction performance evaluation method;case verification showed that the prediction results of combined optimization model have characteristics of minor error,higher precision and better stability.The water traffic accident prediction model based on combined optimization method could enrich the research methods of water traffic accident prediction,and effectively help reducing the risk of water traffic accidents,avoiding water traffic accidents and improving the safety management and decision-making levels of maritime authorities. |
