Relative Risk Estimation Under Multiple Matching
针对罕见病回顾性研究中的多重匹配设计，提出一种非迭代的相对风险估计量，其渐近效率与最大似然估计相当。
Under multiple matching, relative risk from a rare disease is estimated in a retrospective design. Maximum likelihood estimation is carried out using an unconditional probability model. In addition, this model is used to derive a new estimator of relative risk which is noniterative but is as efficient, asymptotically, as the maximum likelihood estimator.