向量法关联规则挖掘在冠心病诊断中的应用

Application of Association Rule Mining Using Vector Method in Coronary Heart Disease Diagnoses

针对传统关联规则频繁项集生成效率较低的问题,提出一种改进的基于向量法的数据关联规则挖掘算法。该算法只需扫描一次事务数据库即可完成布尔矩阵的转换,通过向量运算完成频繁项集的查找,减少候选频繁项集的生成。在冠心病中医诊断中的应用结果表明,该算法可有效提取冠心病中医辨证规则。

Aiming at the problem that the common algorithm often suffers high-complex computation problem, this paper presents an improved association rule algorithm based on vector method, which can map the database into a Boolean matrix by scanning the database only one time. Meanwhile, the frequent itemset can be generated by using simple vector operation. As a result, the number of the frequent items generated by using proposed approach decreases sharply. Experimental results on coronary heart disease data set, including comparisons with the common Apriori approach, illustrate the effectiveness of the proposed algorithm.