中药药对是联系方剂和中药的重要环节。虽然传统中医理论对药对中药味的组成方式有多角度的论述，但相比论著丰富的方剂理论还不够系统。本文分别采用标准关联规则发现Apriori算法以及改进多数据库计算方法，对从历代药对文献中收集整理得到的625个药对347味药中包括性味、归经、功效等共49个属性形成的数据库进行挖掘研究，并对两种方法得到的结果进行比较。结果表明，本文提出的新方法更适用于分析药物间的关联规则，集中度更好，更易于发现有价值的关联规则。

The TCM Crude Drug Pair is an important link node between TCM drugs and formulae. In the traditional TCM theory, there are many descriptions in multi-aspects. However, they seem still scattered compared with formulae theories. In this work, the association rule mining method was applied to reveal the associations between the structures of the crude drug pairs. The standard and enhanced Apriori algorithms were applied and run on the crude drug pair database including 625 drug pairs, 347 drugs and 49 properties. The results from the two algorithms were compared and analyzed. The new enhanced Apriori algorithm was found to be more suitable for the association rule discovery between drugs in the drug pair or formulae.

江苏省高校自然科学研究计划重大基础研究项目（06KJA36022）：四物汤及类方治疗妇科血瘀证的配伍规律与物质基础研究，负责人：段金廒；国家自然科学基金项目（30902015）：方剂图形化技术导向下的方剂效应成分群发现方法的构建，负责人：尚尔鑫；江苏省中医药局科研项目（HZ07048）：中药性味归经配伍规律的可视化方法研究，负责人：尚尔鑫；江苏省中医药局康缘中医药科技创新基金项目（HZ0801KY）：特征性关联方药数据库与知识挖掘，负责人：范欣生。