We study the problem of learning a Bayesian Network structure from data using an Integer Programming approach. We study the existing approaches, an in particular some recent works that formulate the problem as an Integer Programming model. By discussing some weaknesses of the existing approaches, we propose an alternative solution, based on a statistical sparsification of the search space. Results show how our approach can lead to promising results, especially for large networks

An Integer Programming approach to Bayesian Network Structure Learning

Franzin, Alberto
2014/2015

Abstract

We study the problem of learning a Bayesian Network structure from data using an Integer Programming approach. We study the existing approaches, an in particular some recent works that formulate the problem as an Integer Programming model. By discussing some weaknesses of the existing approaches, we propose an alternative solution, based on a statistical sparsification of the search space. Results show how our approach can lead to promising results, especially for large networks
2014-04-15
bayesian, networks, optimization, machine learning, structure learning, integer programming
File in questo prodotto:
File Dimensione Formato  
tesi_franzin.pdf

accesso aperto

Dimensione 1.18 MB
Formato Adobe PDF
1.18 MB Adobe PDF Visualizza/Apri

The text of this website © Università degli studi di Padova. Full Text are published under a non-exclusive license. Metadata are under a CC0 License

Utilizza questo identificativo per citare o creare un link a questo documento: https://hdl.handle.net/20.500.12608/18283