A Two-Step Method of Estimation for Non-Linear Mixed-Effects Models

The main goal of this paper is to propose a two-step method for the estimation parameters in non-linear mixed-effects models. A first-step estimate θ˜ vector θ obtained by solving equations, with working covariance matrix as identity matrix. It shown that consistent. If, furthermore, we have an estimated matrix, V^, θ˜, second-step estimator θ^ can be optimal equations. maintains asymptotic optimality. We establish consistency and normality proposed estimators. Simulation results show improvement over θ˜. Furthermore, provide variance σ2 using moments; also assess empirical performance. Finally, three real-data examples are considered.