基于主题模型的中医辨证关系探索

主题模型因能够有效捕获离散数据间的语义关系和挖掘隐含的主题,而在语义分析、文本分类、信息检索等领域得到广泛的应用。在“证候-症状-用药”的辨证诊疗模式中,主题模型能够挖掘中医多类实体间的复杂语义关系。基于Link-LDA模型,提出了Multi-Link-LDA模型。利用这两种模型,在不同中医诊疗数据集上进行示范性分析,结果显示主题模型能够较好地捕获中医的诊疗规律。

Exploration on the dialectical relationship of TCM based on topic model

Topic model is widely used in semantic analysis, text analysis and information retrieval as it can effectively capture the semantic relationship and mine the latent topics in scattered data. Topic model can mine the complex semantic relationship between variant entities in traditional Chinese medicine according to the "syndrome-symptom-medication" model for diagnosis and treatment by an overall analysis of symptoms and signs. The Multi-Link-LDA model was proposed based on the Link-LDA model. The demonstrative analysis of different datasets for diagnosis and treatment in traditional Chinese medicine based on the two models showed that topic model can effectively capture the rules of diagnosis and treatment in traditional Chinese medicine.