Repository logo
 
No Thumbnail Available
Publication

Exact and heuristic algorithms for variable selection: Extended Leaps and Bounds

Use this identifier to reference this record.
Name:Description:Size:Format: 
012009 - Silva - Extended Leaps and Bounds.pdf129.26 KBAdobe PDF Download

Advisor(s)

Abstract(s)

An implementation of enhanced versions of the classical Leaps and Bounds algorithm for variable selection is provided. Features of this implementation include: (i) The availability of general routines capable of handling many different statistical methodologies and comparison criteria. (ii) Routines designed for exact and heuristic searches. (iii) The possibility of dealing with problems with more variables than observations. The implementation is supplied in two different ways: i) as a C++ library with abstract classes that can be specialized to different problems and criteria. ii) as a console application ready to be applied to searches according to some of the most important comparison criteria proposed to date. The code of the C++ library and console application described here, can be freely obtained by sending an email to the author

Description

Keywords

Variable Selection Algorithms All-Subsets Heuristics

Citation

Silva, A. P. D. (2009). Exact and heuristic algorithms for variable selection: Extended Leaps and Bounds. Working Papers: Economics. N.º 1, 25 p.

Research Projects

Organizational Units

Journal Issue

Publisher

CC License