''On the existence of feature trees on random trees with different distance measures''.

FLESIA, ANA GEORGINA

Balneario Solis. Uruguay. 8 al 10 de noviembre 2007.

Conferencia; IV ERPEM Encuentro regional de Probabilidad y Estadística Matemática; 2007

CENRE (Centro Regional de Probabilidad y Estadística Matemática), Centro de Matemática de la Facultad de Ciencias de la Universidad de la República, la Universidad de San Andrés y el Instituto do Milênio para o Avanço Global da Matematica.

GOAL: Selection of important nodes and“features trees” in supervised and unsupervised classification. This is a correlate of the problem of selection of most informative variables that help to explain a classification method with high-dimensional data. METHODOLOGY:First Approach:  "blinding" uninformative branches of trees, Second Approach: family of distances that take into account notions of probability of each node.MOTIVATION: A strong motivation for this work is to derive a way of distinguish  when two samples of sequences have been produced by different  VLMC models with he same context tree.