基于随机森林和灰色关联法的目标威胁评估  

Target threat assessment based on random forest and grey relational analysis

作　　者：夏鑫 李宁[1] 刘鹏[1] 梁君 XIA Xin;LI Ning;LIU Peng;LIANG Jun(General Department of Information and Electronics,China Ship Research and Development Academy,Beijing 100020,China)

机构地区：[1]中国舰船研究院信息电子技术部,北京100020

出　　处：《舰船科学技术》2024年第13期162-166,共5页Ship Science and Technology

摘　　要：针对目前目标威胁评估中存在的赋权方法无法准确反映指标贡献以及威胁等级划分过于主观化的问题,提出一种新的目标威胁评估模型。该模型融合了随机森林赋权和灰色关联法,通过使用有限的作战指标数据,成功对不同目标的威胁程度进行排名。本文首先建立一个随机森林模型,通过分析模型中各指标与威胁值之间的关系,确定各指标的权重。接着,基于灰色关联法构建了一个威胁等级评估模型,以确保资源分配过程更加客观和合理。结果显示,这种混合评估方法表现出有效性和合理性。In response to the current issues in target threat assessment,where existing weighting methods fail to accurately reflect the contribution of indicators and threat levels are excessively subjective,a novel target threat assessment model is proposed.This model integrates Random Forest weighting and Grey Relational Analysis,successfully ranking the threat levels of different targets using limited operational indicator data.Specifically,an initial Random Forest model is established to analyze the relationships between indicators and threat values,determining the weights of each indicator.Subsequently,a threat level assessment model is constructed based on Grey Relational Analysis,ensuring a more objective and rational allocation of resources.The results demonstrate the effectiveness and rationality of this hybrid assessment approach.

关 键 词：随机森林 威胁等级 评估 

分 类 号：U676[交通运输工程—船舶及航道工程]