摘要
目的讨论零膨胀模型在计数资料中的应用。方法应用零膨胀负二项模型分析冠心病患者心肌缺血节段数的影响因素。结果零膨胀负二项模型logit部分结果表明没有家族史、年龄越小、左室收缩末容积越小的患者发生心肌节段缺血的可能性较小;负二项部分结果表明有糖尿病史、有冠心病家族史、左室收缩末容积越大的患者发生心肌缺血节段数越多。结论当计数资料存在过多零计数时,应用零过多模型能够得到很好的拟合效果。
Objective To discuss the application of zero-inflated models(ZIM) in count data.Methods Analyse the impacting factors about the segment number of myocardial ischemia applying zero-inflated negative binomial model.Results The part of logit showed no family history of Coronary Heart Disease,younger,smaller left ventricular end-systolic volume were the factors that impacted whether or not myocardial ischemia occurred.The part of negative binomial displayed that the increasing segment number of myocardial ische...
出处
《中国卫生统计》
CSCD
北大核心
2008年第5期464-466,共3页
Chinese Journal of Health Statistics
基金
山西省自然基金项目(20051091)