摘要
本文对一种新型中药降脂灵片的药效进行因果分析。实验数据样本量小且是混合变量类型,传统的统计方法难以处理,本文采用图模型的方法建立一个链图模型,直观地刻画了该药对反映机体抗氧化能力和血脂水平的4个指标的因果影响。
This paper made causal inferences for the efficacy of Jiailgzhiling tablet, a new type of Chinese medicine. Because of the small sample size and the mixed variable types, traditional statistical approaches can not give reliable results. In this paper, a chain graphical model was developed to describe the causal effect of Jiangzhiling on four variables which measured antioxidant capacity and blood fat level.
出处
《数理统计与管理》
CSSCI
北大核心
2007年第4期741-746,共6页
Journal of Applied Statistics and Management
基金
国家自然科学基金NSFC(10571070)
中科院系统所许国志博士后基金资助项目
关键词
链图
条件Gauss分布
条件独立性
模型选择
chain graph
conditional Gaussian distribution
conditional independence
model Selection
