基于因果模型的中医证候研究探讨

Causal Model Based Analysis of TCM Syndrome Studies

中医学的的基本特点之一是辨证论治,传统的辨证方法往往带有主观性和模糊性。由于证候的确定缺乏客观、定量的研究方法,以致许多诊治经验难以重复,因此,探索中医证候的客观化和量化研究方法是一个重要课题。近年来,在统计学、临床流行病学和数据挖掘领域的研究中,因果推断越来越受到重视。目前,有若干种因果推断的模型,其中结构方程模型和贝叶斯网络模型被越来越多地应用在中医证候研究中,其可以客观、量化研究中医证候。本文主要介绍、分析这两种模型的原理、构建方法、在中医证候研究中的应用情况。

Dialectical treatment is a basic feature of TCM. Unfortunately, traditional dialectical methods are always characterized with subjectivity and ambiguity, which leaves many treatment experiences unrepeatable. To overcome the unneeded part of traditional methods, both objective and quantitative research methods are introduced to differentiate syndromes. Recently, causal inference has paid more attention to statistics, epidemiology, and data mining. Its two models： structural equation model and Bayesian network, constitute an objective and quantitative approach of TCM study. The two models have found increasing applications in the study of syndrome differentiation. In this article, authors analyze the principles and methods of building a structural equation model and the Bayesian network, and present a review of applications of the two models in syndrome differentiation.