基于XGBoost的船舶仿冒行为监测方法  

Monitoring Method of Ship Counterfeiting Based on XGBoost

作　　者：隋远 段然 白正 SUI Yuan;DUAN Ran;BAI Zheng(Nanjing LES Cybersecurity and Information Technology Research Institute Co.Ltd.,Nanjing 210014,China)

机构地区：[1]南京莱斯网信技术研究院有限公司,南京210014

出　　处：《指挥信息系统与技术》2022年第5期60-65,79,共7页Command Information System and Technology

基　　金：2020年交通运输部战略规划政策研究(2020-14-2)资助项目。

摘　　要：针对水上安全监管中无法有效识别船舶类型的问题,提出了一种基于XGBoost的船舶仿冒行为监测方法。首先,利用船舶历史航迹数据计算得到不同类型船舶的航迹特征;然后,利用XGBoost算法对船舶航迹特征进行训练得到船舶分类模型;最后,提出了一种船舶类型自动检测流程,实现了船舶仿冒行为监测。试验结果表明,该方法对船舶类型分类判断具有较高的分类准确率、较快的训练收敛速度和较高的分类判断效率。Aimed at the problem of ship types cannot be identified effectively in water safety supervision,a ship counterfeiting monitoring method of based on XGBoost is proposed.Firstly,the track characteristics of different types of ships are calculated by using the historical track data of ships.Then,the ship classification model is obtained by using XGBoost algorithm to train the ship track characteristics.Finally,an automatic detection process of ship type is proposed,and the monitoring of ship faking behavior is realized.The experimental results show that this method has the higher classification accuracy,the faster training convergence speed and the higher classification judgment efficiency.

关 键 词：XGBoost 仿冒行为 监测方法 

分 类 号：U675.3[交通运输工程—船舶及航道工程]