Study On Combined Prediction Method Of Water Traffic Accident Based On IOWA Operator

With the sustainable and stable development of waterage,Shipping economy has become an important branch of national economy and played a decisive role in the process of national modernization drive.Although the rapid development of waterage brings some economic benefits,more potential safety hazards in water traffic occurred along with the development.Water traffic accident has not only caused heavy economic losses but also seriously threaten the life of people.So we need to effectively predict water traffic accident in a reasonable way which can holistically control the trend of water traffic accident.It is of great realistic significance to reduce the risk of accident and improve security of water traffic.This paper focuses on the analysis and prediction of the factors affecting the water traffic accidents.By Combination Forecasting and Induced Ordered Weighted Averaging,water traffic Combined Forecasting Models are built,which are also verified by means of living examples.Compared with the error index values of combined forecasting models,the accuracy and applicability of the models are proved.The paper mainly research as below:Firstly,An integrated DEMATEL,ISM and ANP method is established to analyze the influence factors of water traffic accidents.Because water traffic system is a dynamic,uncertain and random complex system,we can built a water traffic accident influencing factor hierarchical structural model with an integrated approach to turn the intra-system relationship from a complex factor relationship into a clear and simple hierarchical one.Furthermore,the internal operating mechanism of influencing factors is clearly defined to obtain the key influencing factors.Then,by summary and analysis of the applicability of the water traffic accident forecasting method,we can built two kinds of Single variable Forecasting Models.One is multivariable grey model MGM(1,4)which can take into account multiple relative factors,the other is SARIMA Time Series Forecasting Model.Through consideration of the inaccuracy between the Background Values of traditional multivariable grey models,we optimized the model.According to the results of optimized MGM(1,4)and SARIMA,we built water traffic accident Combination Forecasting Model that is also based on Induced Ordered Weighted Averaging(IOWA).By use of Markov chain the accuracy state of the out-of-sample data is deduced qualitatively.So the Weight coefficient of the model is verified as well.At last,referred to the statistical data in Tianjin waterage as history data,we separately forecast water traffic accident of in-sample 2008—2017 and out-of-sample 2018—2020 by both the Single variable Forecasting Model and Combination Forecasting Model.Compared with the error index values of the Forecasting Models and Fitted Forecasted values and actual values,it proved that Combination Forecasting Model can effectively reduce the prediction error.The building of Combination Forecasting Model provides new research directions to water traffic accident forecasting.