基于关联规则的中医证候分类模型应用研究

Application Research of the TCM Syndrome Classification Model Based on Association Rules

中医证候分类研究是中医辨证施治研究的核心问题之一,挖掘病例样本中不同证候间的关联关系、处方信息与证候间的关联关系具有重要的研究价值与临床应用意义。本研究结合中医证候分类中关联分析的特征选择需要,以四诊信息标准化为基础,提出了一种新的病例样本量化方法;以量化后的病例样本为挖掘分析的多维数据集,基于关联规则优化的FP-Growth算法构建了中医证候关联分析模型。模型在采集到的1499个高血压样本数据上进行了分析,证候关联挖掘结果,“神疲乏力”“精神萎靡”关联性最强;证方挖掘结果“杜仲、构藤、天麻”是“肝阳上亢”证候的有效处方,实验结果充分验证了提出方法的有效性。

The research of TCM syndrome classification is one of the core issues of TCM syndrome differentiation and treatment.It is of great research value and clinical significance to mine the correlation between different syndromes,the relationship between prescription information and syndromes in case samples.In this paper,a new method of case sample quantification is proposed based on the standardization of four diagnostic information and the feature selection of association analysis in TCM syndrome classification.The quantized case sample is used as a multidimensional data set.We build a TCM syndrome correlation analysis model based on Frequent Pattern Tree(FP-Tree),which is improved Association Rules algorithms.The model was analyzed on the collected data of 1,499 hypertension samples.The results of syndrome correlation mining showed that the correlation between"faint fatigue"and"dementia"was the strongest;the mining results of the evidence"Duzhong,Gouteng,Tianma"were the effective prescription of the syndrome of“Ganyangshangkang”.This experimental results fully verify the effectiveness of our proposed method.