摘要
针对我国跨境邮寄新业态呈现爆发式发展,为实现业务“管得住、通得快”,为实现对进出境邮件风险源的检测和智能预警,构建风险监控、风险预警、查验处置的一体化体系,降低风险邮寄包裹中动植物、有害生物、生化风险物等高风险源的漏检问题,提升跨境进出境邮件监管效能。本文参考大数据模型的构建方法,采用逻辑回归模型对跨境邮寄物的申报数据、历史查验数据、各类检测监管设备产生的过程信息进行大数据分析,并对潜在风险包裹进行风险预警。建立的逻辑回归模型能对风险包裹数据进行分类,比对预警结果和真实结果能达到较高的准确率。建立的邮件风险源监测和智能预警具有科学性和有效性,适用于应对进出口跨境邮寄件的快速预警,达到高效的监测效果。
With the explosive development of the new cross-border mail business in China,in order to achieve goals of"fast clearance but manageable",accomplishes the detection and intelligent early warning of inbound and outbound mail risk sources,builds an integrated system of risk monitoring,risk early warning,inspection and disposal,reduces the missed detection of high-risk sources such as animals and plants,harmful organisms,and biochemical risk substances in risky parcels,so as to improve the supervision efficiency of cross-border inbound and outbound mail.Referring to the construction method of the big data model,using the logistic regression model to carry out big data analysis on the declaration data of cross-border mail,historical inspection data,and process information generated by various inspection and supervision equipment,and to send out risk warning for potentially risky packages.The established logistic regression model can analyze and classify the risk package data,and can achieve a high accuracy rate by comparing the early warning results with the real results.The established mail risk source monitoring and intelligent early warning are scientific and effective in application,and are suitable for rapid early warning of import and export cross-border mail,so as to achieve efficient monitoring results.
作者
罗琴涛
张宗平
罗宇平
胡琳子
梁军峰
梁志明
LUO Qintao;ZHANG Zongping;LUO Yuping;HU Linzi;LIANG Junfeng;LIANG Zhiming(Guangzhou Customs Information Center,Guangzhou Guangdong,510623,China;China Electronic Port Data Center Guangzhou Branch;Foshan Customs Comprehensive Technical Service Center)
出处
《质量安全与检验检测》
2022年第3期60-63,共4页
QUALITY SAFETY INSPECTION AND TESTING
基金
跨境邮寄物中风险源在线可视化识别与处置技术(2018YFC0809200)。
关键词
风险预警
风险挖掘模型
逻辑回归
邻近模型
Risk Pre-Warning
Risk Mining Model
Logistic Regression
K-Nearest Neighbor