The Partial Linear Variable Coefficient Model Based On Empirical Likelihood Of Statistical Diagnosis

As the promotion of a variable coefficient model and partial linear model, partially linear varying-coefficient partially linear model has been widely used. This paper is based on empirical likelihood for varying-coefficient partially linear model for statistical diagnosis. First of all, the estimators are given for the unknown parameter and coefficient function, their asymptotic behaviors are proved; then, we develop diagnostic measures for empirical likelihood of general estimating equations. Our diagnostics include case-deletion, local influence measures and pseudo-residuals. Finally, stochastic simulation and data analysis are given to illustrate the validity of statistical diagnostic measures.