基于数据挖掘的围绝经期综合征中医证候分类算法分析

摘要

Objective To establish the optimum syndrome classification method by using the technology of modern TCM diagnosis and artificial intelligence analysis method for menopausal syndrome differentiation of TCM. Methods Diagnostic information of menopausal syndrome patients was collected and syndromes were classified according to TCM syndrome differentiation standard. Three kinds of common data mining classification algorithm, Bayesian network, K-nearest neighbors and support vector machine, were used for analysis on information data of the four methods of diagnosis of menopausal syndrome.Results The time, classification accuracy, coverage rate and margin curve of establishing TCM syndrome model by the three kinds of algorithm methods under the circumstances of same training and data. The influence of the number of training samples of 3 kinds of algorithm methods was analyzed, and the model established by the three kinds of algorithms was evaluated.Conclusion Bayesian network algorithm is better than the other two methods in the menopausal syndrome classification effect.%目的：采用现代中医诊断技术结合人工智能分析方法进行围绝经期综合征中医辨证研究，以期建立最佳证候分类方法。方法门诊收集围绝经期综合征患者四诊信息，按照中医辨证标准进行证型分类，采用贝叶斯网络算法、K最近邻算法、支持向量机算法3种常用数据挖掘分类算法对围绝经期综合征四诊信息数据进行分析。结果分别得出在相同训练、测试样本数据下3种算法建立围绝经期综合征中医证候模型所需时间、分类准确性、覆盖率及margin曲线，分析了训练样本数量对3种算法的影响，并对3种算法所建立模型进行了评价。结论在围绝经期综合征证候分类效果方面，贝叶斯网络算法优于其他2种方法。