一种基于带权无向图的中医方剂频繁项集挖掘算法

摘要

根据中医方剂数据的特点,将频繁项集发现算法应用到中医方剂的研究中,挖掘治疗消渴病胃火炽盛症型的方剂中不同药对之间的关联规则以及核心药物组合,提出一种基于带权的无向频繁项集图的挖掘算法.该算法可以快速挖掘频繁k-items(k≥3),并随之快速回溯频繁项集所对应的方剂,从而完成了中医方剂数据特点的快速数据挖掘.通过实验表明,该算法避免产生大量候选项集,提高了中医方剂数据挖掘效率,对完成中医消渴病方剂信息的用药规律分析具有重要意义.%According to the characteristics of traditional Chinese medicine prescription data, the frequent item sets discovery algorithm is used to the study of TCM prescriptions, and mine the association rule and core drug combination of the different drugs in the prescriptions for the treatment of Emaciation-Thirst disease with stomach fire flaming.In this paper, frequent item sets mining algorithm based on weighted undirected graph is proposed.The proposed algorithm can rapidly mine frequent k-items, and quickly backtrack to the prescription corresponding to frequent item sets, thus accomplishing the rapid data mining of TCM prescription data.Experiments show that this algorithm avoids generating a large number of candidate items and improves the data mining efficiency of TCM prescription, which is of great significance to analyze the law of drug use in the prescriptions of Emaciation-Thirst disease.