肥胖症的患病因素研究与预测  

Research and Analysis on Risk Factors of Obesity

作　　者：黄萃琳[1] 陆余楚[1] 

机构地区：[1]上海大学理学院数学系,上海200444

出　　处：《应用数学与计算数学学报》2007年第1期82-88,共7页Communication on Applied Mathematics and Computation

基　　金：复旦-瑞士再保险研究基金.

摘　　要：肥胖症是一组常见的代谢症候群,其发病率在中国逐年上升.影响肥胖症的因素很多,本文研究的是年龄、肥胖症家族史、吸烟时间、吸烟数量、饮酒时间、饮酒频率、饮酒数量和户外活动.另外,由于肥胖症通常会有并发症,所以,还附加了高血压,冠心病,糖尿病,高血脂这四个疾病的相关指标,总共16个指标.本文首先用单因子Logistic回归挑选出与肥胖症患病相关性较大的因素,然后用主成分分析方法消除因素间的共线性,最后用标准化自变量的Logistic回归模型将这些因素对肥胖症患病影响的重要性程度进行排序,同时拟合出患病概率的预测模型.Obesity is one of the most universal metabolic diseases and its incidence of disease is increasing gradually in China. Many factors lead to Obesity. They are namely age, Obesity family heredity, smoking time and quantity, drinking time, frequency and quantity, out-door activity. In addition, Obesity always tends to complications, so it is necessary to include some factors of Hypertension, Coronary heart disease, Diabetes mellitus and Hyperlipidemia in the thesis, totally sixteen factors. In this method, single factor logistic regression is used to select the more important factors, thus by the means of main composition analysis, linearity connection among factors is removed, finally rank these factors by the individual degree of their influence to Obesity by standardized independent variable, and imitate the sick probability model by logistic regression.

关 键 词：肥胖症 主成分分析 LOGISTIC回归 

分 类 号：R589.2[医药卫生—内分泌]