Nonparametric Estimation of the Regression Function in an Errors-in-Variables Model

Abstract: We consider the regression model with errors-in-variables where we observe

n

i.i.d. copies of

(Y, Z)

satisfying

Y = f (X) + x i, Z = X + s i g m a e p s i l o n

, involving independent and unobserved random variables

X, x i, e p s i l o n

. The density

g

of

X

is unknown, whereas the density of

s i g m a e p s i l o n

is completely known. Using the observations

(Y_{i}, Z_{i})

,

i = 1, ..., n

, we propose an estimator of the regression function

f

, built as the ratio of two penalized minimum contrast estimators of

e l l = f g

and

g

, without any prior knowledge on their smoothness. We prove that its

m a t h b b L_{2}

-risk on a compact set is bounded by the sum of the two

m a t h b b L_{2} (m a t h b b R)

-risks of the estimators of

e l l

and

g

, and give the rate of convergence of such estimators for various smoothness classes for

e l l

and

g

, when the errors

e p s i l o n

are either ordinary smooth or super smooth. The resulting rate is optimal in a minimax sense in all cases where lower bounds are available.

This page was built for publication: Nonparametric Estimation of the Regression Function in an Errors-in-Variables Model