基于关联规则的轻工品涉税检验数据挖掘方法

Association Rule Mining for Tariff Related Inspection Data of Imported Light Industrial Consumer Products

进口轻工消费品的货值金额快速增长,类别与数量也大幅增加,快速准确完成现场涉税监管,对一线人员是个挑战。借鉴关联规则在其他领域的应用,提出了一种进口轻工消费品涉税检验数据挖掘方法。首先,从税则归类、检验方式和检验时间三个维度对进口轻工消费品涉税检验数据进行概念分层,并作为扩展信息加入到原始数据中;其次,对进口轻工消费品涉税检验数据中的多值型信息项进行多值映射,转化为多个布尔型信息项;最后,结合多值映射后同类信息项不能同时出现的特点,改进Apriori算法挖掘关联规则。对某海关进口轻工消费品涉税检验数据的挖掘结果表明,所提方法能够发现强关联的信息项,可为一线人员提供重点关注清单,辅助提升现场监管效率。

The amount of impor ted light industrial consumer goods have increased rapidly, and the category and quantity have also increased significantly. It is a challenge for front-line personnel to quickly and accurately complete the on-site tariff related supervision. Based on the application of association rules in other fields, a tariff related inspection data mining method of imported light industrial consumer goods was proposed. Firstly, the tariff related inspection data of imported light industrial consumer goods was conceptually stratified from three dimensions: tariff classification,inspection method and inspection time, and added to the original data as extended information.Secondly, the multi value information items in the tariff related inspection data of imported light industrial consumer goods were mapped and transformed into multiple Boolean information items.Finally, considering that information items with the same type cannot appear at the same time after multi value mapping, Apriori algorithm was improved to mine association rules. The mining results of tariff related inspection data of imported light industrial consumer goods from a certain customs showed that the proposed method can find strongly related information items, provide front-line personnel with a list of key concerns, and help improve the efficiency of on-site supervision.