ATE模型中匹配的可重复性问题分析  

Analysis of Matching with/without Replacement in ATE Model

作　　者：郭凯旋[1] 

机构地区：[1]上海财经大学数学系,上海200439

出　　处：《数理医药学杂志》2005年第4期289-292,共4页Journal of Mathematical Medicine

基　　金：国家自然科学基金(基金号:10171059);"面向21世纪教育振兴行动计划"专项资金(基金号:200218)资助

摘　　要：“治疗的平均效果”问题起源于医学统计学领域,其涉及的模型现在广泛应用于医学、生物学、经济学等学科。匹配是解决这一问题的重要方法,从理论证明和计算机模拟两个方面分析了可重复匹配与不可重复匹配两种方法的优劣及各自特点。'Average treatment effects'problem originated from the field of medical statistics. The model that has been mentioned is now being widely applied in the sciences of iatrology,biology, economics, and so on. Matching is an important method for solving the problem. In this paper, both the advantage and disadvantage of matching with/without replacement has been analyzed through theoretical proof and computer simulation.

关 键 词：匹配 平均治疗效果 可重复匹配 不可重复匹配 

分 类 号：O21[理学—概率论与数理统计] R311[理学—数学]