Journal of the American Statistical Association · 1974 · 15 citations · 0 references
ComputingComputer ScienceAlgorithmsOptimizationUncertainty QuantificationEstimation TheoryIterative SolversNumerical AlgorithmsParameter EstimationEfficient EstimationJoint Location EstimationNewton-raphson OptimizationRelative Efficiency
Abstract It is shown that for the estimation of the location parameter of a Cauchy distribution, for which the scale parameter is known, k-spacings corresponding to the maximum asymptotic relative efficiency (ARE) are asymmetric when k = 4m − 1; this result is in contrast to Chan's assumption that optimal k-spacings are always symmetric. Accurate solutions for the maximal ARE and optimal k-spacings obtained by a combination of dynamic programming and Newton-Raphson iterations have been tabulated for k = 3, 7 and 11. In the joint estimation of the location and scale parameters the maximum joint asymptotic relative efficiency is achieved by the equi-probable k-spacing λi = i/(k + 1).