Statistics and Its Interface

Volume 6 (2013)

Number 4

Coordinate great circle descent algorithm with application to single-index models

Pages: 511 – 518



Yichao Wu (Department of Statistics, North Carolina State University, Raleigh, N.C., U.S.A.)

Peng Zeng (Department of Mathematics and Statistics, Auburn University, Auburn, Alabama, U.S.A.)


Coordinate descent algorithm has been widely used to solve high dimensional optimization problems with a nondifferentiable objective function recently. To provide theoretical justification, Tseng (2001) showed that it leads to a stationary point when the non-differentiable part of the objective function is separable. Motivated by the single index model, we consider optimization problems with a unit-norm constraint in this article. Because of this unit-norm constraint, the coordinate descent algorithm cannot be applied. In addition, non-separability of the non-differentiable part of the objective function makes the result of Tseng (2001) not directly applicable. In this paper, we propose a novel coordinate great circle descent algorithm to solve this family of optimization problems. The validity of the algorithm is justified both theoretically and via simulation studies. We also use the Boston housing data to illustrate this algorithm by applying it to fit single-index models.


constrained optimization, coordinate descent algorithm, penalization, single-index model, unit-norm constraint

2010 Mathematics Subject Classification

Primary 62F10, 62H12. Secondary 62P05.

Full Text (PDF format)