摘要： 在幂集理论的基础上，引入约束幂集概念，提出一种依赖于约束幂集的频繁邻近类别集(NCS)挖掘算法。该算法采用计算约束幂集映射的方法，生成候选频繁NCS并计算支持数，使其能避免冗余候选项的产生以及减少对数据库的重复扫描次数。实验结果表明，该算法在挖掘约束频繁NCS时比现有挖掘算法更快速有效。

关键词: 邻近类别集, 约束条件, 幂集映射, 约束幂集, 空间关联规则, 空间数据挖掘

Abstract: This paper proposes a constraint power set concept based on power set theory, and proposes an algorithm of mining frequent Neighboring Class Set(NCS) depending on constrain power set. The algorithm uses computing constraint power set mapping to generate candidate frequent NCS and computes its support count. It not only avoids generating redundant candidate, but also reduces calculation amount of repeated scanning database. Experimental result indicates that the algorithm is faster and more efficient than present mining algorithm when extracting constraint frequent NCS.

Key words: Neighboring Class Set(NCS), constraint condition, power set mapping, constraint power set, spatial association rule, spatial data mining

中图分类号: 

• N945