Very Simple Classification Rules Perform Well on Most Commonly Used Datasets

By Robert Craig Holte. 1993

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Evaluation & Benchmarks

Citations in the corpus

Interpretable Machine Learning: Fundamental Principles and 10 Grand Challenges (2021, background): “It has been known for a very long time that very simple models perform surprisingly well for tabular data (Holte, 1993).”
Please Stop Explaining Black Box Models for High Stakes Decisions (2018, background): “This ﬁeld dates back to the early 90’s at least [see Freitas, 2014, Holte, 1993], and there are a huge number of papers on interpretable ML in various ﬁelds, which do not have the word “interpretable" or “explainable" in the title, as the recent papers do.”
Learning Certifiably Optimal Rule Lists (2017, background): “Interpretable Models: There is a growing interest in interpretable (transparent, compre-hensible) models because of their societal importance (see R¨uping, 2006; Bratko, 1997; Dawes, 1979; Vellido et al., 2012; Giraud-Carrier, 1998; Holte, 1993; Shmueli, 2010; Huys-mans et al., 2011; Freitas, 2014).”
Scalable Bayesian Rule Lists (2016, neutral): “Interpretability has long since been a fundamental topic in artiﬁcial intelligence (see R¨uping, 2006; Bratko, 1997; Dawes, 1979; Vellido, Mart´ın-Guerrero, & Lisboa, 2012; Giraud-Carrier, 1998; Holte, 1993; Shmueli, 2010; Huysmans, Dejaeger, Mues, Vanthienen, & Baesens, 2011; Freitas, 2014).”
Box drawings for learning with imbalanced data (2014, neutral): “This, however, is in contrast with other views (e.g., (Holte 1993)) that for most common datasets, simple rules exists and we should explore them.”
Or's of And's for Interpretable Classification, with Application to Context-Aware Recommender Systems (2015, methodology): “Some of these methods, like CBA, CPAR and CMAR [10, 11, 26, 51] still suffer from a huge number of patterns and do not yield interpretable classifiers, yet it is well-known that for many domains, the space of good predictive models is often large enough to include very simple models [20].”
Methods and Models for Interpretable Linear Classification (2014, background): “There is also classic work supporting the idea that simple interpretable models can be accurate [41], and that domain expertise can be used to construct accurate models [19].”
Finding a short and accurate decision rule in disjunctive normal form by exhaustive search (2010, methodology): “• 1R (Holte 1993): is a baseline learner that induces a classifier based on one single attribute.”
Supersparse Linear Integer Models for Interpretable Classification (2013, background): “There is classic work supporting the idea that simple, interpretable models have the capability to perform well [32].”
A statistical approach to rule learning (2006, methodology): “As argued by Holte (1993), decision stumps are representationally very simple, but often feature only slightly worse predictive accuracy than more sophisticated approaches.”
Building Interpretable Classifiers with Rules using Bayesian Analysis (2012, background): “This was hypothesized over a decade ago [18], however, we now are starting to have the computational tools to truly test this hypothesis.”
An Interpretable Stroke Prediction Model using Rules and Bayesian Analysis (2013, unknown): “Call the rule˜r i .”
Benchmarking Least Squares Support Vector Machine Classifiers (2004, methodology): “…classifiers including the SVM classifier, decision tree based algorithms, statistical algorithms and instance based learning methods (Aha & Kibler, 1991; Breiman et al., 1984; Duda & Hart, 1973; Holte, 1993; John & Langley, 1995; Lim, Loh, & Shih, 2000; Ripley, 1996; Witten & Frank, 2000).”
A bias/variance decomposition for models using collective inference (2008, background): “On the other hand, very simple models can sometimes outperform complex models due to decreased variance, albeit with higher bias (Holte 1993).”