摘要
基于FAS算法,设计并实现了一个基于最近挖掘结果的数据挖掘系统AR_Miner,包括数据预处理、频繁集初始计算、频繁集更新计算、频繁集选择、关联规则生成5个部分,实现了关联规则挖掘的可视化和生成结果按“支持度-可信度”形式的可视化,为基于频繁集的交互式挖掘提供了方便、友好的界面。对历史的稽查数据中纳税人采用的主要违法违章手段之间的关联关系进行了数据挖掘,得到了一些合理的知识,提高了对稽查历史数据的管理水平,为税务稽查提供了有力的决策支持。
A FAS-based data mining system AR Miner with the latest mining results are designed. It consists of five parts： data preprocessing, initial calculation of frequent itemsets, update calculation of frequent itemsets, choice of frequent itemsets and generation of association rules. It visualizes generated results in support-confidence form, and provides an easily accessible and user-friendly interface for interactive mining based on frequent itemsets. The paper uses it into mining the association rules of tax offence of the historical data, receives some reasonable knowledge, and increases the level of inspection historical data management to provide a strong tax reform policy support.
出处
《计算机工程》
CAS
CSCD
北大核心
2007年第14期283-284,F0003,共3页
Computer Engineering
关键词
数据挖掘
关联规则算法
FAS算法
data mining
algorithm of association rules
FAS algorithm
