Strong Consistency of a Class of Estimators in Partial Linear Model for Negative Associated Samples

Consider the heteroscedastic regression model: y_i=x_iβ+g（t_i）+σ_ie_i, 1≤i≤n, where σ_i²=f（u_i）. Here the design points （x_i,t_i,u_i） are known and nonrandom, g and f are unknown functions, β is an unknown parameter to be estimated, and the errors e_i are negatively associated random variables. For the least squares estimator \widehat\beta_n_n and the weighted least squares estimator \bar\beta_n_n of β, we establish their strong consistency under suitable conditions.