基于主题模型982例心脑合病患者证候特征研究

目的：探索冠心病心绞痛合并动脉粥样硬化性血栓性脑梗死（心脑合病）的证候分类及其特征。方法：应用临床流行病学调查的横断面研究方法，在中国11家三甲医院收集982例心脑合病患者的症状、体征、舌脉等中医四诊信息并建立数据库，采用频数统计及主题模型等数据挖掘方法，分析心脑合病的证候。结果：常见的四诊信息：肢体不遂、胸闷、乏力、头晕、身体困重、神疲、少气懒言、胸痛、肢体麻木、舌下脉络紫色等；挖掘出瘀热内蕴、阴虚内热、痰湿阻络、气血亏虚、痰瘀互结、脾肾阳虚、肝郁脾虚、气虚血瘀、瘀血阻滞、肝肾阴虚10类证候，并初步确定了各证候的诊断要点。结论：心脑合病中医证候可分10型，运用主题模型的无监督数据分析方法研究证候分类及证候特征有一定的科学性、客观性和实用性。

Study on Syndrome Characteristic of 982 Angina Pectoris Complicated with Atherosclerotic Thrombotic Cerebral Infarction Cases Based on Topic Model

This study explored the syndrome characteristic of angina pectoris complicated with atherosclerotic thrombotic cerebral infarction (XNHB for short). Clinical epidemiological research methods were applied. Information of 982 XNHB cases from 11 triple-A hospitals in China were collected to build database, including general condition, symptoms and signs, tongue pictures, pulse conditions, and etc. Data mining techniques such as frequency statistics and topic model were used in the mining of syndrome characteristic of XNHB. The results showed that the high occurrence frequency of information from traditional Chinese medicine (TCM) four diagnostic methods in turn were limbs disable, chest distress, hypodynamia, dizziness, physical heaviness, languor, laziness to speak, chest pain, limb numbness, purple color of subglossal collateral vessels and so on. There were 10 TCM syndrome patterns, including stasis and heat accumulation, yin deficiency and inner heat, accumulation of phlegm-damp in collaterals, deficiency of qi and blood, accumulation of phlegm and blood stasis, yang deficiency of spleen and kidney, liver depression and spleen deficiency, qi deficiency and blood stasis, blockage of blood stasis, yin deficiency of liver and kidney. The diagnostic point of each XNHB syndrome pattern was established preliminarily. It was concluded that there were 10 TCM syndrome patterns of XNHB. It was scientific, objective and practical to study the syndrome classification and characteristics of syndrome in topic model.