【作者】 曾涛； 向勇； 包小源； 刘胤田； 蒋永光；

【Author】 ZENG Tao~1 XIANG Yong~2 BAO Xiao-yuan~1 LIU Yin-tian~2 JIANG Yong-guang~3 (College of Computer and Information Engineering,Tianjin Normal University,Tianjin 300387,China)1 (School of Computer Science,Sichuan University,Chengdu 610065,China)2 (Chengdu University of Traditional Chinese Medicine,Chengdu 610075,China)3

【机构】 天津师范大学计算机与信息工程学院； 四川大学计算机学院； 成都中医药大学药学院；

【摘要】 中医方剂内药物配组的作用受方剂语境影响,药物搭配方式不同,功效则不同,药物配伍原理和机制尚不明确,是中医古方研究的重要内容。发现频繁药组,并对其进行功效标注可以多粒度、多维度、多方向,更深入地研究中药方剂药物配伍规律,对中医药学科研究具有重要意义,提出并形式化描述了该问题,讨论了相关性质,设计并实现了可直接进行多表关联分析的基于逆索引的2阶段频繁模式挖掘方法,讨论了算法特性。通过在真实古方数据集上较全面深入的实验研究,表明新算法可行、效果良好,合理及良好标注率可达71%左右,所有结果均对中医药物配伍及功效关系研究有参考价值。更多还原

【Abstract】 Medicine Groups(MG) in Traditional Chinese Medicine(TCM) prescription have various medicinal effects under the influence of prescription semantic context without straightforward mechanism.MG problem is very significant for TCM research.In order to study it multi-granularly,multi-dimensionally,and multi-directionally,techniques of Frequent Medicine Group mining and corresponding medicinal Effect Labeling(FMGEL) were introduced, formal concepts were proposed and analyzed;a new algorithm,based on 2-phase frequent pattern mining mode and inverted-index techniques,were designed and implemented,which can make association analysis on multi-relation; property of new algorithm were discussed.Experimental results on real TCM prescription data show that new method is feasible and instructive,71%of FMGEL results are reasonable or fine,and all of FMGEL results are valuable.更多还原