A class of linear regression parameter estimators constructed by nonparametric estimation (Q1094776)

The parameters of the multiple linear regression function \(\alpha (x)=E(Y| X=x)=X\beta\) may be estimated from a sample \(\{(X_ i,Y_ i)\), \(i=1,...,n\}\) by minimizing the functional \[ {\hat \psi}(\beta)= \int({\hat\alpha}_ n(x)-x\beta)^ 2\hat f_ n(x)dx, \] where \(\hat f_ n(x)\) is a nonparametric kernel-estimator of the unknown density function f(x) of the random variable X and \({\hat\alpha}_ n(x)= \sum^{n}_{i=1} Y_ iI_{\{X_ i\}}(x)\) is a nonparametric estimator of \(\alpha(x)\). The strong consistency and asymptotic normality of this estimator are proved. This class of estimators includes generalized ridge regression estimators as a special case.