From 13c5d898053de79deb29adc43a2e5438551e7021 Mon Sep 17 00:00:00 2001 From: Shruthi Chari Date: Wed, 5 Jun 2024 21:01:55 -0400 Subject: [PATCH] Adding metrics for rules - need to support them in Metaexplainer next --- Ontologies/v3/explanation-ontology.owl | 35 ++++++++++++++++++++++++++ 1 file changed, 35 insertions(+) diff --git a/Ontologies/v3/explanation-ontology.owl b/Ontologies/v3/explanation-ontology.owl index e9989ef..c95b0d5 100644 --- a/Ontologies/v3/explanation-ontology.owl +++ b/Ontologies/v3/explanation-ontology.owl @@ -1034,6 +1034,18 @@ + + + + + Lakkaraju, H., Kamar, E., Caruana, R., & Leskovec, J. (2017). Interpretable & explorable approximations of black box models. arXiv preprint arXiv:1707.01154. + Average Rule Length + The average number of antecedents, connected with the AND operator, of the rules contained in each ruleset +Lakkaraju et al. (2016); Wu et al. (2018). a, represents the number of antecedents of the ith rule and R = |A| the number of rules + + + + @@ -2122,6 +2134,17 @@ understanding and knowledge [(McNeil and Krajack, 2008) of how the world works, + + + + + Saad, E. W., & Wunsch II, D. C. (2007). Neural network explanation using inversion. Neural networks, 20(1), 78-93. + Fidelity + Ratio of input instances on which the predictions of model and rules agree (f) over total instances + + + + @@ -2697,6 +2720,18 @@ Source: O'Connor, K. J., & Ammen, S. (2012). Play therapy treatment pla + + + + + + + + + + + + Rules