基于改进SCGM(1,1)_(C)模型的海上交通事故量预测

Forecast of Traffic Accident at Sea with Improved SCGM(1,1)_(C) Model

为提高海上交通事故量的预测精度,以单因数系统云灰色预测模型(System Cloud Grey Model,SCGM(1,1)_(C))为基础,提出基于马尔可夫预测理论修正和粒子群算法优化的改进SCGM(1,1)_(C)模型。阐述SCGM(1,1)_(C)预测模型的建模过程;结合SCGM(1,1)_(C)与马尔科夫预测理论的优点构造马尔科夫SCGM(1,1)_(C)预测模型;利用粒子群算法优化马尔科夫状态区间白化系数,得到经过2次修正的改进SCGM(1,1)_(C)预测模型;以2005—2019年海上交通事故实际数据为样本,使用这3种模型分别进行预测计算,并作相应预测值的拟合曲线图。结果表明:改进SCGM(1,1)_(C)模型的预测精度和拟合性较另外2种模型有大幅度的提高,为海上交通事故量预测问题研究提供一种新的方法。

An SCGM(1,1)_(C)(System Cloud Grey Model,SCGM(1.1);)Model is combined with Markov prediction and Particle swarm operation to improve the accuracy of traffic accident forecast at sea.The Markov SCGM(1,1)_(C)prediction model is built,incorporating the advantages of both Markov prediction and traditional SCGM(1,1)_(C).Whitening transformation coefficient of Markov State region is optimized by particle swarm optimization.This makes a double correction of the SCGM(1,1)_(C)prediction model.The traffic accident data in the period of 2005—2019 is calculated with 3 model types respectively and fitting curves are drawn,which proves the advantage of the improved prediction model.