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
-
摘要: 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。Abstract: An SCGM（1,1）C（System Cloud Grey Model, SCGM（1.1）C） 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.
下载: