摘要： 传统的关联规则挖掘算法易形成大量频繁项目集，不适用于异构环境下海量交通数据的挖掘。为此，提出基于层次梯度且无候选项分析的协同数据挖掘算法。采用挖掘主题数据库和层次梯度构建层次业务数据库，逐层深度挖掘局部频繁项。利用弱化熵模型对频繁项主题数据库进行数据分析，并产生关联规则。实验结果表明，该算法适用于无候选项支持的协同挖掘。

关键词: 协同数据挖掘, 关联规则, 层次梯度, 层次业务数据库

Abstract: The classic data mining algorithm produces a lot of frequent-item set, which is not applied to the massive data mining in Intelligent Transportation System(ITS). This paper proposes an algorithm based on level grads without candidate items analysis that is used for computing association rules under the heterogeneous environment. It uses the concept of both level grads and mining topic transaction databases forming the level transaction database and mining the local frequent-item. The main-node uses the concept of weakly-entropy to abstract some association rules. Simulation results show that this algorithm has better performance in collaborative mining without candidate support.

Key words: collaborative data mining, association rule, level grads, level transaction database

中图分类号: 

• TP311