Methods and Models for Interpretable Linear Classification

By Berk Ustun, Cynthia Rudin. 2014

46 citations recorded by the source corpus.

We present a new comprehensive approach to create accurate and interpretable linear classication models through mixed-integer programming. Unlike existing approaches to linear classication, our approach can produce models that incorporate a wide range of interpretability-related qualities, and achieve a precise, understandable and pareto-optimal balance between accuracy and interpretability. We demonstrate the value of our approach by using it to train scoring systems, M-of-N rule tables, and \personalized" classication models for applications in medicine, marketing, and crime. In addition, we

Subfields: Artificial Intelligence, Statistics and Probability, Safety, Robustness & Interpretability

How researchers see this paper

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

In Artificial Intelligence

In Statistics and Probability

In Safety, Robustness & Interpretability

Citations in the corpus

TreeView: Peeking into Deep Neural Networks Via Feature-Space Partitioning (2016, background): “There have been deliberate efforts to develop models that are interpretable in the recent past [4], [2], [5] .”
The Bayesian Case Model: A Generative Approach for Case-Based Reasoning and Prototype Classification (2014, methodology): “In our view, there are at least three widely known types of interpretable models: sparse linear classifiers ([30, 8, 31]); discretization methods, such as decision trees and decision lists (e.”
Interpretable classification models for recidivism prediction (2015, methodology): “To this end, we use a powerful new machine learning method known as a Supersparse Linear Integer Model (SLIM, see Ustun et al. 2013, Ustun and Rudin 2014, 2015).”
Learning Credible Models (2017, background): “Credibility is closely related to interpretability, which has been actively explored in the literature [10, 17, 19, 24, 28, 30].”
Association of an Electroencephalography-Based Risk Score With Seizure Probability in Hospitalized Patients (2017, neutral): “0 1 2 3 4 5 >6 Probable risk of Sz, %b 5 12 27 50 73 88 >95 Actual prevalence of Sz, % (95% CI)c 3 (2-3) 12 (10-13) 34 (31-37) 52 (46-57) 71 (63-78) 84 (71-99) 92 (77-100)”
Learning Classification Models of Cognitive Conditions from Subtle Behaviors in the Digital Clock Drawing Test (2015, methodology): “We use a recently developed framework, Supersparse Linear Integer Models (SLIM) (Ustun and Rudin 2015; Ustun et al. 2013), and introduce a simple metric to prioritize more understandable features, enabling us to build interpretable linear models.”
Explaining Classifiers with Causal Concept Effect (CaCE) (2019, background): “Many approaches have been proposed to tackle this challenge: for example by developing a model with interpretable components built-in [8, 16] or by building post-training interpretability methods [13, 14, 4, 2, 1, 9].”
Precinct or Prejudice? Understanding Racial Disparities in New York City's Stop-and-Frisk Policy (2016, background): “…these difficulties, we draw on a large body of work which has found that simple, transparent and interpretable heuristics often work as well as complex statistical models [Czerlinski, Gigerenzer and Goldstein (1999), Gigerenzer and Goldstein (1996), Lovie and Lovie (1986), Ustun and Rudin (2014)].”
A Semiquantitative Group Testing Approach for Learning Interpretable Clinical Prediction Rules (2015, methodology): “Recent machine learning research has attempted to learn clinical prediction rules that generalize accurately from large-scale electronic health record data rather than relying on manual development [4], [5].”
Interactive and interpretable machine learning models for human machine collaboration (2015, unknown): “In my view, there are at least three widely known types of interpretable models: sparse linear classifiers ([126, 32, 129]); discretization methods, such as decision trees and decision lists (e.”
Interpretability of machine learning models and representations: an introduction (2016, background): “An explanatory model "relates attributes to outcomes in a clear, informative, and meaningful way" [17].”
A Survey on the Explainability of Supervised Machine Learning (2020, methodology): “Supersparse linear integer models (SLIM) is a method used for making scoring systems more interpretable (Ustun & Rudin, 2014).”
Towards Automatic Concept-based Explanations (2019, background): “This work is focused on post-training explanation methods - explaining an already trained model instead of building an inherently explainable model [42, 19, 40].”
Analysis of Explainers of Black Box Deep Neural Networks for Computer Vision: A Survey (2019, methodology): “We give some examples of explaining approaches that can be placed in the last point: A global ante-hoc method for tabulator data is Bayesian Rule List (BRL) [66], [67].”