Ecole d'ingénieur et centre de recherche en Sciences du numérique

LIBRE: Learning interpretable boolean rule ensembles

Mita, Graziano; Papotti, Paolo; Filippone, Maurizio; Michiardi, Pietro

AISTATS 2020, 23rd International Conference on Artificial Intelligence and Statistics, 3-5 June 2020, Palermo, Sicily, Italy

We present a novel method – libre– to learn an interpretable classifier, which materializes as a set of Boolean rules. libre uses an ensemble of bottom-up, weak learners operating on a random subset of features, which allows for the learning of rules that generalize well on unseen data even in imbalanced settings. Weak learners are combined with a simple union so that the final ensemble is also interpretable. Experimental results indicate that libre efficiently strikes the right balance between prediction accuracy, which is competitive with black box methods, and interpretability, which is often superior to alternative methods from the literature.

Document Arxiv Bibtex

Titre:LIBRE: Learning interpretable boolean rule ensembles
Type:Conférence
Langue:English
Ville:Palermo
Pays:ITALIE
Date:
Département:Data Science
Eurecom ref:6125
Copyright: © EURECOM. Personal use of this material is permitted. The definitive version of this paper was published in AISTATS 2020, 23rd International Conference on Artificial Intelligence and Statistics, 3-5 June 2020, Palermo, Sicily, Italy and is available at :
Bibtex: @inproceedings{EURECOM+6125, year = {2020}, title = {{LIBRE}: {L}earning interpretable boolean rule ensembles}, author = {{M}ita, {G}raziano and {P}apotti, {P}aolo and {F}ilippone, {M}aurizio and {M}ichiardi, {P}ietro}, booktitle = {{AISTATS} 2020, 23rd {I}nternational {C}onference on {A}rtificial {I}ntelligence and {S}tatistics, 3-5 {J}une 2020, {P}alermo, {S}icily, {I}taly}, address = {{P}alermo, {ITALIE}}, month = {06}, url = {http://www.eurecom.fr/publication/6125} }
Voir aussi: