Jackknife Method for Variance Components Estimation of Partial EIV Model

Abstract

To further improve the quality of estimated values based on variance component estimation of the partial errors-in-variables (EIV) model, the jackknife resampling method is introduced in this paper. Focusing on the bias of variance component estimation and combining with the jackknife method, bias calculation and bias correction are performed. Two schemes for parameter estimation are identified, and detailed calculation steps and the whole procedure are given. The jackknife method for variance component estimation of the partial EIV model is evaluated. Meanwhile, these two new algorithms are applied to the straight-line fitting model, space-line fitting model, and plane coordinate transformation model. As shown in the experimental estimation results, both methods proposed can obtain more accurate estimated values than the traditional variance component estimation method, and the method with bias correction can obtain the optimal parameter estimates. The case studies demonstrate the effectiveness and reliability of the proposed procedure, which extends the theory of the jackknife method in parameter estimation and provides resampling insight to further investigate variance component estimation.