Corpus ID: 208075751

LIBRE: Learning Interpretable Boolean Rule Ensembles

@article{Mita2020LIBRELI,
  title={LIBRE: Learning Interpretable Boolean Rule Ensembles},
  author={Graziano Mita and Paolo Papotti and Maurizio Filippone and Pietro Michiardi},
  journal={ArXiv},
  year={2020},
  volume={abs/1911.06537}
}
  • Graziano Mita, Paolo Papotti, +1 author Pietro Michiardi
  • Published 2020
  • Computer Science, Mathematics
  • ArXiv
  • 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… CONTINUE READING

    Figures, Tables, and Topics from this paper.

    References

    Publications referenced by this paper.
    SHOWING 1-10 OF 40 REFERENCES

    A Bayesian Framework for Learning Rule Sets for Interpretable Classification

    VIEW 3 EXCERPTS

    An Imbalanced Data Rule Learner

    VIEW 1 EXCERPT

    Fast Effective Rule Induction

    VIEW 3 EXCERPTS