Estimation by Simulation
将McFadden的模拟矩方法推广到一般估计问题中，通过模拟偏差校正来近似有效得分，并讨论了偏差校正与矩近似之间的效率权衡。
We extend McFadden's Method of Simulated Mo- ments to approximating efficient estimators in general estima- tion problems. The general approach applies a simulated bias correction to an approximation of the efficient score. We discuss a general trade-off in estimator inefficiency between bias correction and moment approximation. I. Introduction T HE method of moments (MOM) for estima- tion of an unknown parameter vector 0 with true value 00 often begins with the moment con- ditions E{W(y - E(yIX,00))} = 0 requiring orthogonality between a weighting ma- trix W = W(X, 0) and a residual vector y - E(y I X, 0), where y is a random vector which has a conditional expectation E(y I X, 0) and X is a vector of covariates. In its simplest form, the method derives an estimator 0 from analog sam- ple moments: EN{W(y - E(yIX,0))} -EyI 0)~ 0 = Nx -W( Yn -E(ylXn o) n=1I