摘 要: 海关监管部门在风险布控的过程中,需要风险分析人员依据经验对各种商品的风险进行人工分类。本文用Logistic回归模型、决策树模型、随机森林模型等几种的分类模型优化风险布控过程,通过将报关数据输入分类模型得到特定商品的风险评估结果,辅助风险分析人员做出正确判断。通过实验验证这种智能化的方法能够有效克服人工风险布控中的不足,完成智能化风险布控,进一步维护国门口岸安全。 |
关键词: 大数据;机器学习;分类;风险布控 |
中图分类号: TP183
文献标识码: A
|
基金项目: 2020年度四川省科技厅创新基地(平台)和人才计划《四川对外贸易数据分析与风险防控》(2020JDR0330). |
|
Application of Multiple Classification Models in Intelligent Customs Risk Prevention and Control |
JIN Jin1,WANG Zhenggang2,LIU Wei2,WU Jiamin1, LI Bo1
|
( 1.Chengdu Neusoft University, Chengdu 611844, China; 2.Chengdu Customs of the People 's Republic of China, Chengdu 610041, China )
jinjin@nsu.edu.cn; wangzgxs@outlook.com; 45711577@qq.com; WuJiamin@nsu.edu.cn; li-bo@nsu.edu.cn
|
Abstract: As part of the process of risk prevention and control, the customs supervision department requires risk analysts to manually classify risks of various commodities based on their experiences of risk management. This paper uses several classification models, such as logistic model, decision tree model, and random forest model, to optimize risk control process. Risk assessment results of specific commodities can be obtained by inputting customs declaration data into the classification models. Thus the results assist risk analysts to make correct judgments. The proposed model is verified through experiments. The results show that it is an intelligent method and can effectively overcome the shortcomings in manual risk control, complete the intelligent risk control, and further maintain the security of national ports. |
Keywords: big data; machine learning; classification; risk prevention and control |