Probabilistic conditional independence: a similarity-based measure and its application to causal network learning

A new definition for similarity between possibility distributions is introduced and discussed as a basis for detecting dependence between variables by measuring the similarity degree of their respective distributions. This new definition is used to detect conditional independence relations in possib...

Descripción completa

Detalles Bibliográficos
Autores: Sangüesa i Sole, Ramon|||0000-0002-0998-4179, Cabós, Joan, Cortés García, Claudio Ulises|||0000-0003-0192-3096
Tipo de recurso: informe técnico
Fecha de publicación:1996
País:España
Institución:Universitat Politècnica de Catalunya (UPC)
Repositorio:UPCommons. Portal del coneixement obert de la UPC
Idioma:inglés
OAI Identifier:oai:upcommons.upc.edu:2117/84564
Acceso en línea:https://hdl.handle.net/2117/84564
Access Level:acceso abierto
Palabra clave:Similarity
Possibility distributions
POSSCAUSE
Àrees temàtiques de la UPC::Informàtica::Intel·ligència artificial
Descripción
Sumario:A new definition for similarity between possibility distributions is introduced and discussed as a basis for detecting dependence between variables by measuring the similarity degree of their respective distributions. This new definition is used to detect conditional independence relations in possibility distributions derived from data. This is the basis for a new hybrid algorithm for recovering possibilistic causal networks. The algorithm POSSCAUSE is presented and its applications discussed and compared with analogous developments in possibilistic and probabilistic causal networks learning.