隐树模型几个关键指标的辨证意义  

Key indexes' meaning of the syndrome differentiation on the latent tree models

作　　者：袁世宏[1] 张连文[2] 王天芳[3] 赵燕[3] 王庆国[3] 

机构地区：[1]山西职工医学院,太原030012 [2]香港科技大学,中国香港 [3]北京中医药大学,北京100029

出　　处：《中华中医药杂志》2012年第5期1241-1244,共4页China Journal of Traditional Chinese Medicine and Pharmacy

基　　金：国家重点基础研究发展计划(973计划)(No.2011CB505105);香港研究资助局项目(No.622408);北京中医药大学创新团队项目(No.2011-CXTD-08)~~

摘　　要：数据挖掘的各种算法模型能否真正量化地解释证候,关键是该模型所产生的有关指标能否合理地量化解释证和一组症状的关系。文章以抑郁症临床流行病学调查数据构筑的证候隐树模型为例子,分析了该模型所产生的互信息、累积互信息、信息覆盖度、条件概率等几个指标的辨证意义。认为:互信息可作为确定一组与某证有密切关联的症状的依据;累积互信息和互信息数值之间的比较可以把握症状提供给证的信息,从而判断症状的诊断价值;信息覆盖度可以考察与某证相关联的一组症状中究竟有多少症状、或有哪些症状就足可以把握该证的基本特征;而条件概率则可以通过该证关联的一组症状所表现出的变化来定量地刻画这个特征。Whether a kind of algorithm model can quantitatively interpret syndrome of traditional Chinese medicine, the key is whether some of the indexes from the model can reasonably and quantitatively explain the relationship between the syndrome and symptoms. And this paper take the latent tree model of syndrome, Which be constructed by the data from the depression, as an example, to analyze the differentiation syndrome meaning of the indexes, such as mutual information （MI）, cumulate mutual information （CMI）, information rate （IR）, conditional probability. Subsequently, the MI may be as a important basis for confirming the association between a syndrome and a group of symptoms; and meanwhile, according to the comparison of the MI and CMI, we could be obtained the syndrome information given by symptoms and estimate the diagnostic value of the symptoms. And then the IR can be used to estimate how many symptoms and which symptoms, which are close relevance to one of syndrome, can be took as a confirming of the intrinsic character of the syndrome, and finally, conditional probability could contribute to a better revealing the character of a syndrome quantitatively by the changing that a groups of symptoms relevance to a syndrome had happened on different condition.

关 键 词：隐树模型 证 互信息 累积互信息 信息覆盖度 条件概率 

分 类 号：R241[医药卫生—中医诊断学] R277.7[医药卫生—中医临床基础]