1: \begin{abstract}

2: The Dantzig selector \citep{candes2007dantzig} is a popular

3: $\ell^1$-regularization method for variable selection and estimation

4: in linear regression.  We present a very weak geometric condition on

5: the observed predictors which is related to parallelism and, when satisfied,

6: ensures the uniqueness of Dantzig selector

7: estimators. The condition holds with

8: probability 1, if the predictors are drawn from a continuous

9: distribution.  We discuss the necessity of this condition for

10: uniqueness and also provide a closely

11: related condition which ensures uniqueness of lasso estimators

12: \citep{tibshirani1996regression}. Large sample

13: asymptotics for the Dantzig selector, i.e. almost sure convergence and the asymptotic

14: distribution, follow directly from our uniqueness results and a

15: continuity argument.  The limiting

16: distribution of the Dantzig selector is generally non-normal. Though our

17: asymptotic results require that the number of predictors is fixed

18: (similar to \citep{knight2000asymptotics}), our uniqueness results are

19: valid for an arbitrary number of predictors and observations.

20: \end{abstract}