Abstract: In order to resolve the mining problem of the quantitative association rule, a concept of similar attributes set matrix is proposed. In the meanwhile, a new algorithm of the quantitative association rules(Qarmasm) is proposed based on this concept. The algorithm does not need to expend attribution of tradition, and it has higher efficiency of reducing. It can produce pre-frequent itemsets and get support directly, and find frequent itemsets efficiently. The paper analyzes the complexity of algorithm and validates its validity, gives the application and contrast to indicate that this algorithm is much better than the traditional algorithm.

Key words: similar attributes set matrix, frequent pattern, association rule, data mining

摘要： 为解决多值属性的关联规则挖掘问题给出相似属性集合矩阵的概念，提出一种新的多值关联规则挖掘算法——Qarmasm算法。该算法无须扩展事务属性，约简效率高，能够直接生成候选频繁项集，求出其支持度，有效地发现频繁项。给出算法的描述及其复杂性分析。与经典算法的对比表明，该算法具有明显的优势。

关键词: 相似属性集合矩阵, 频繁模式, 关联规则, 数据挖掘

CLC Number: 

• TP312