Traditionally, differentiation of syndromes of Traditional Chinese Medicine (TCM) mainly depends on the information obtained from four diagnosis methods. Now many physicochemical parameters are available in clinic. Th...Traditionally, differentiation of syndromes of Traditional Chinese Medicine (TCM) mainly depends on the information obtained from four diagnosis methods. Now many physicochemical parameters are available in clinic. There exists great correlation between TCM syndromes and physicochemical parameters. The objective of the paper is to analyze the correlation between TCM syndromes and physicochemical parameters quantitatively. Correlation analysis has been widely studied and many analysis methods have been developed. Mutual information based on entropy can measure arbitrary dependence between variables. It has been applied to many kinds of fields, especially to pattern recognition. But most works are restricted to discrete variables and little work has been done to study the relation between discrete and continuous variables. A novel algorithm is proposed to calculate the mutual information between discrete and continuous variables. It is used to analyze the correlation between TCM syndromes and physicochemical parameters.展开更多
基金The research has been supported by National Basic Research Program of China(973 Grant 2003CB517106) MOSTProjects (No.2004DFB02100)
文摘Traditionally, differentiation of syndromes of Traditional Chinese Medicine (TCM) mainly depends on the information obtained from four diagnosis methods. Now many physicochemical parameters are available in clinic. There exists great correlation between TCM syndromes and physicochemical parameters. The objective of the paper is to analyze the correlation between TCM syndromes and physicochemical parameters quantitatively. Correlation analysis has been widely studied and many analysis methods have been developed. Mutual information based on entropy can measure arbitrary dependence between variables. It has been applied to many kinds of fields, especially to pattern recognition. But most works are restricted to discrete variables and little work has been done to study the relation between discrete and continuous variables. A novel algorithm is proposed to calculate the mutual information between discrete and continuous variables. It is used to analyze the correlation between TCM syndromes and physicochemical parameters.
