中医处方的关联分析与频次统计方法异同探析

Exploration the Similarities and Differences of Correlation Analysis and Frequency Statistics Methods of TCM Prescription

目的:通过比较中医处方的关联分析与频次统计方法在中医处方中应用的异同,从而获取较好解读处方关联分析的方式。方法:以妇科医案中的435个痛经处方为研究对象,分别进行了关联分析与频次统计,从计算公式及结果方面进行了两种方法异同的比较。结果:关联分析中两个测度值—支持度和置信度代表含义差异较大。通过支持度筛选的计算结果与频次统计类似,而通过置信度筛选的计算结果,频次统计方法难以企及。结论:置信度高而支持度低的药对才是处方关联分析的特色结果,是值得关注的数据挖掘结果。

Objective：To obtain the best method for understanding the result of formula data by comparison the similarities and differ- ences between method of correlation analysis and frequency analysis applying to TCM prescription. Methods ： The 435 prescriptions of dysmenorrheal in the gynecology medical case as the object of study, and were carried out both correlation analysis and frequency analy- sis. The similarities and differences of the two methods were compared from formula and result. Results ： The degree of support and confi- dence in correlation analysis represents different meaning. The calculation result of correlation analysis is similar to frequency statistics by the degree of support screening,but the calculation result of correlation analysis was quite better than frequency statistics by the de- gree of confidence screening. Conclusion： In the Chinese herb result data mining result, the Chinese herb couples with the high degree of support and low degree of confidence are worth to be pay more attention.