基于特征优选和SVM的船舶航行事故致因分析

在“海运强国”战略和建设“海上丝绸之路”的大背景下,我国对船舶航行安全提出更高的要求。本文基于特征优选和支持向量机模型,挖掘出船舶航行事故的致因并分析各个因素对事故的影响程度。首先,通过文本挖掘和相关性分析对输入特征进行优选,筛选出航行事故责任船舶与其他船舶存在明显差异的因素作为航行事故致因。然后,构建基于SVM的船舶航行事故识别模型,并通过交叉验证及群体智能优化算法选择模型的最佳参数组合,得到最优的分类模型。最后,利用递归特征消除算法将上述致因对事故的影响程度进行排序和筛选,挖掘出事故的关键致因。通过广东省的水上交通事故实例验证模型的有效性,结果表明:本模型(正确度为90.1%)较传统单一的SVM模型(正确度为75.0%)具有更高的精度。研究结果可为减少船舶航行事故提供有效的科学建议。

图加一条边后的带宽和

令BS(G,f)=∑|f(u)—f(v)|,其中f为V(G)→{1,2,…,|V(G)|}的双射,并称BS(G)=min BS(G,f)为图G的带宽和.讨论顶点数为n的简单图G加上一条边e∈E(G)后,带宽和BS(G+e)与BS(G)的关系,得其关系式BS(G)+1≤BS(G+e)≤BS(G)+n-1.并证明此不等式中等号可取到,即存在图G_1和G_2使得BS(G_1+e)=BS(G_1)+1,BS(G_2+e)=BS(G_2)+n-1.

Ergodicity of Bandwidth and Cutwidth on Families of Graphs and Trees

Bandwidth,cutwidth,cyclic bandwidth,bandwidth sum and cyclic bandwidth sum are well-known indices about optimal labeling of graphs applied in VLSI design,network communications,and other areas involving the graph layout.To design the graphs with the given indices,we need to study the ergodicity.Let F be a set of graphs under consideration andφan integer-valued function defined on F,namely,φis an index,such as bandwidth and cutwidth.If there exists a graph G∈F such thatφ(G)=x for any integer x in the interval[a,b],where a and b are the minimum and maximum ofφon F,respectively,thenφis said to have ergodicity on F.Let Gnbe the set of simple connected graphs with order n and Tnthe set of trees with order n.In this paper,we investigate the ergodicity of bandwidth,cutwidth,cyclic bandwidth,the bandwidth sum and cyclic bandwidth sum on Tn and Gn.