[Statlist] Séminaires de Statistique - ISTAT

[KONDYLIS Atanassios]

Séminaires de Statistique 
Institut de Statistique, Université de Neuchâtel 
Espace de l'Europe 4, Neuchâtel
http://www2.unine.ch/statistics



Mardi 13 juin 2006 à 11h00 
-----------------------------------------------
Ph.D. Cristian Gatu, University of Neuchâtel, Switzerland

Title:  Branch-and-bound strategies for deriving the best-subset regression models 

Abstract : An efficient branch-and-bound algorithm for computing the best-subset
regression models is described. The algorithm is based on a regression tree structure 
that generates all possible subset models. The computational burden is reduced by 
pruning the non-optimal subtrees. The criteria used in identifying the best subsets are 
based on functions of the residual sum of squares. Strategies and heuristics which improve 
the computational performance of the proposed algorithm are investigated. Computational 
results and experiments on random and real data are presented and analyzed.

For abstracts and more info on the Statistical Seminars in the 
Institute of Statistics, see http://www2.unine.ch/statistics/page9123.html