Numerical Methods to Estimate Parameters in Quantile Regression  

Hector Manuel Mora                           

 

 

Abstract

Quantile regression is a nondifferentiable convex optimization problem. We compare three classical numerical methods, two of them based on linear optimization, and the cutting plane method. We compare them by their required memory and computing time.

 

Key words: Quantile regression, Linear programming, Nondifferentiable optimization, Cutting planes.

 

PDF (Spanish)