Events and Meetings of Italian Statistical Society, Statistics and Demography: the Legacy of Corrado Gini

Font Size: 
A proposal for learning Bayesian networks from categorical variables
Flaminia Musella

Last modified: 2015-09-05

Abstract


Bayesian networks are graphical models that represent the joint distribution of a set of variables using directed acyclic graphs. If the domain knowledge is available the graph can be manually built, otherwise the network has to be inferred from data by using suitable learning algorithms. In this paper, we deal with a constraint-based method to perform Bayesian networks structural learning in the presence of mixed nominal-ordinal categorical variables. We propose a revision of the PC algorithm by using nonparametric tests automatically selected according to the variables typology.


Full Text: PDF