Combining multiple directed graphical representations into a single probabilistic model

• Enrique Castillo ^[1] ; José Manuel Gutiérrez ^[1] ; Ali S. Hadi ^[2]
  1. [1] Universidad de Cantabria

     Universidad de Cantabria

     Santander, España

     
  2. [2] Cornell University

     Cornell University

     City of Ithaca, Estados Unidos

     
• Localización: CAEPIA'97: actas / coord. por Asociación Española de Inteligencia Artificial, Vicente J. Botti Navarro , 1997, ISBN 84-8498-765-5, págs. 645-652
• Idioma: inglés
• Texto completo no disponible (Saber más ...)
• Resumen

  • In this paper we deal with the problem of combining the knowledge of multiple human experts into a single probabilistic model. More specifically, we analyze how to combine the information contained in a set of directed acyclic graphs (a set of casual models) into a single probabilistic structure. In particular, we show that, in cases where there exists an ancestral ordering of the variables compatible with all the graphs, the resulting probabilistic structure is given by a set of conditional probability distributions obtained from the topology of the intersection network.