当前位置: 首页 > 期刊 > 《第四军医大学学报》 > 2005年第3期
编号:10965319
广义可加模型共曲线性及其在空气污染问题研究中的应用
http://www.100md.com 《第四军医大学学报》 2005年第3期
统计学,方法,,统计学,方法;共曲线性;模型,统计学;空气污染,0引言,1方法原理,2实例分析,3讨论
     Concurvity in generalized additive models in study of air pollution

    JIA Bin1,WANG Tong1,WANG LinNa1,CHEN ChangSheng2

    1Department of Health Statistics,School of Public Health, Shanxi Medical University, Taiyuan 030001, China, 2Department of Health Statistics,School of Preventive Medicine,Fourth Military Medical University,Xian 710033,China

    【Abstract】 AIM: To explore the effect of concurvity of generalized additive models (GAMs) in the study of air pollution and human health and to diminish the influence of concurvity. METHODS: We employed the nonparametric conditional bootstrap method to diminish the influence of concurvity. RESULTS: This method got a better confidence interval. CONCLUSION: The nonparametric conditional bootstrap method can overcome the effect of concurvity in GAMs and get a better result.

    【Keywords】 statistics/methods;concurvity;models, statistical;air pollution

    【摘要】 目的: 针对大气污染与健康关系研究中拟合广义可加模型时的共曲线性现象进行研究并试图减小其影响. 方法: 采用非参数条件自助抽样法来估计模型系数及其可信区间. 结果: 实例数据分析可见自助法得到的结果较好. 结论: 非参数条件自助抽样法是克服广义可加模型中共曲线性影响的有益尝试.

    【关键词】 统计学/方法;共曲线性;模型,统计学;空气污染 ......

您现在查看是摘要页,全文长 10870 字符