长江干线水上交通事故的灰色分析与预测

Grey Analysis and Prediction of Maritime Traffic Accidents on the Trunk Line of Yangtze River

为减少长江海事局辖区水上交通事故的发生,增强水上交通运输的安全,采用灰色系统理论中的邓氏关联理论和改进的灰色关联理论,对2008—2017年长江海事局辖区水上交通事故统计数据进行关联性分析,定量分析水上交通事故的主要关联因素。将不同灰色关联理论分析的结果进行比较分析,给出一种与定性分析吻合度较高的灰色关联分析方法,提高水上交通事故分析的精确度。对不符合灰色系统建模要求的随机序列,引入实用序列缓冲算子,分别构建均值和加权GM(1,1)预测模型,对长江海事局辖区未来水上交通事故进行预测,并对不同预测模型的预测值和实际值进行比对,给出一种精度达到0.7%的短期预测模型,为改善长江海事局辖区的水上交通安全形势提供参考依据。

The analysis is made from 2008 to 2017 statistical data of maritime traffic accidents in the district under Yangtze River Maritime Bureau to find the main factors involved in the accidents. Quantitative analysis is performed by means of both traditional and improved grey relational analysis. By comparing the results from different analysis, a grey relational analysis method of higher degree of consistency with qualitative analysis is chosen for improving the accuracy of analyzing maritime traffic accidents. For non-smooth sequences a practical sequence buffer operator is introduced, which allows the construction of the even grey model and weighted grey model for predicting maritime traffic accidents. A finalized short-term prediction model achieves prediction accuracy of 0.7% and is of a reference for marine safety in the area.