Bayesian Rule Sets for Interpretable Classification

Wang, Tong; Rudin, Cynthia; Velez-Doshi, Finale; Liu, Yimin; Klampfl, Erica; MacNeille, Perry

doi:10.1109/icdm.2016.0171

Cited by 134 publications

(220 citation statements)

References 14 publications

Supporting

Mentioning

209

Contrasting

Order By: Relevance

“…Rule-based systems are probably the most interpretable models because their IF-THEN structure semantically resembles natural language and the way human think [20]. There are also many researches to construct the more interpretable rule-based systems [21,22].…”

Section: A Interpretabilitymentioning

confidence: 99%

Concise Fuzzy System Modeling Integrating Soft Subspace Clustering and Sparse Learning

Deng

Zhang

et al. 2019

IEEE Trans. Fuzzy Syst.

View full text Add to dashboard Cite

show abstract

Section: A Interpretabilitymentioning

confidence: 99%

Concise Fuzzy System Modeling Integrating Soft Subspace Clustering and Sparse Learning

Deng

Zhang

et al. 2019

IEEE Trans. Fuzzy Syst.

View full text Add to dashboard Cite

show abstract

“…such that the error E between the MLP's outputsŷ n and the given targets y n is minimized. In (1), Ω(W) is a so-called regularization term, whose strength can be controlled by a regularization parameter λ ∈ R + . Usually, this term is used to avoid overfitting.…”

Section: A Multi-layer Perceptron (Mlp)mentioning

confidence: 99%

“…One way to enforce so-called explainability of ML models is to create models for which the explainability objective is an essential part of their design. Ante-hoc models are designed to be inherently explainable; examples for this type of models are logistic regression, rule-based systems [1], or decision trees. In contrast, post-hoc explainability refers to adding the explainability objective after training [2], [3].…”

Section: Introductionmentioning

confidence: 99%

Enhancing Decision Tree Based Interpretation of Deep Neural Networks through L1-Orthogonal Regularization

Schaaf

Huber

Maucher

2019

2019 18th IEEE International Conference on Machine Learning and Applications (ICMLA)

View full text Add to dashboard Cite

One obstacle that so far prevents the introduction of machine learning models primarily in critical areas is the lack of explainability. In this work, a practicable approach of gaining explainability of deep artificial neural networks (NN) using an interpretable surrogate model based on decision trees is presented. Simply fitting a decision tree to a trained NN usually leads to unsatisfactory results in terms of accuracy and fidelity. Using L1-orthogonal regularization during training, however, preserves the accuracy of the NN, while it can be closely approximated by small decision trees. Tests with different data sets confirm that L1-orthogonal regularization yields models of lower complexity and at the same time higher fidelity compared to other regularizers.

show abstract

“…1 Current Trends. Current work in inductive rule learning is focused on finding simple rules via optimization (Dash et al, 2018;Wang et al, 2017;Malioutov and Meel, 2018), mostly with the goal that simple rules are more interpretable. However, there is also some evidence that shorter rules are not always more convincing than more complex rules (Fürnkranz et al, 2018;Stecher et al, 2016).…”

Section: Induction Of Predictive Rule Setsmentioning

confidence: 99%

Advances in Machine Learning for the Behavioral Sciences

Kliegr

Bahník

Fürnkranz

2019

American Behavioral Scientist

View full text Add to dashboard Cite

The areas of machine learning and knowledge discovery in databases have considerably matured in recent years. In this article, we briefly review recent developments as well as classical algorithms that stood the test of time. Our goal is to provide a general introduction into different tasks such as learning from tabular data, behavioral data, or textual data, with a particular focus on actual and potential applications in behavioral sciences. The supplemental appendix to the article also provides practical guidance for using the methods by pointing the reader to proven software implementations. The focus is on R, but we also cover some libraries in other programming languages as well as systems with easy-to-use graphical interfaces.

show abstract

Bayesian Rule Sets for Interpretable Classification

Cited by 134 publications

References 14 publications

Concise Fuzzy System Modeling Integrating Soft Subspace Clustering and Sparse Learning

Concise Fuzzy System Modeling Integrating Soft Subspace Clustering and Sparse Learning

Enhancing Decision Tree Based Interpretation of Deep Neural Networks through L1-Orthogonal Regularization

Advances in Machine Learning for the Behavioral Sciences

Contact Info

Product

Resources

About