V. Alabau, J. Andrés,  F. Casacuberta, J. Civera, J. García-Hernández,
A. Giménez, A. Juan, A. Sanchis, and E. Vidal.  The naive Bayes model,
generalisations  and   applications.   In  F.  Pla,   P.  Radeva,  and
J.  Vitrià,  editors, Pattern  Recognition:  Progress, Directions  and
Applications,  pages 162-179.  Centre de  Visió per  Computador, 2006.
ISBN 84-933652-6-2.

The naive  Bayes classification model is a  very simple classification
technique   in   which   pattern    features   are   assumed   to   be
class-conditional independent.   This is the so-called  naive Bayes or
independence  assumption.  In  spite  of being  a strong,  unrealistic
assumption, the naive  Bayes model often provides good  results at low
cost in terms  of model complexity. The Pattern  Recognition and Human
Language  Technology  group  from  the  Universitat  Polit\`ecnica  de
Val\`encia  maintains  an active  research  line  on  this model,  its
generalisations  (mainly  discrete  mixture models)  and  applications
(text classification, word  disambiguation and confidence measures for
speech recognition, etc.).