1: \begin{abstract}

2: Because of the advance in technologies,

3: modern statistical studies often encounter linear models

4: with the number of explanatory variables much larger than

5: the sample size.

6: Estimation and variable selection in these high-dimensional

7: problems with deterministic design points

8: is very different from those in the case

9: of random covariates, due to the

10: identifiability of the high-dimensional

11: regression parameter vector. We show that a reasonable

12: approach is to focus on the projection of the regression parameter vector

13: onto the linear space generated by the design matrix.

14: In this work, we consider the ridge regression estimator of the

15: projection vector

16: and propose to threshold the ridge regression estimator when the

17: projection vector is sparse in the sense that many of its components

18: are small. The proposed estimator has an explicit form and

19: is easy to use in application.

20: Asymptotic properties such as the consistency of

21: variable selection and estimation and the

22: convergence rate of the prediction mean squared error are established

23: under some sparsity conditions on the projection

24: vector. A simulation study is also conducted to examine the

25: performance of the proposed estimator.

26: \end{abstract}