Beta calibration: a well-founded and easily implemented improvement on logistic calibration for binary classifiers

By Meelis Kull, T. S. Filho, Peter A. Flach. 2017

295 citations recorded by the source corpus.

How researchers see this paper

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

Citations in the corpus

Calibrating Deep Convolutional Gaussian Processes (2018, methodology): “The calibration properties of standard classifiers and neural networks have been studied in the literature [17, 23], which has shown that classifiers that use the standard cross-entropy loss are generally well calibrated.”
Aleatoric and Epistemic Uncertainty in Machine Learning: A Tutorial Introduction (2019, methodology): “Examples of such methods include binning (Zadrozny and Elkan, 2001), isotonic regression (Zadrozny and Elkan, 2002), logistic scaling (Platt, 1999) and improvements thereof (Kull et al., 2017), as well as the use of Venn predictors (Johansson et al.”
Beyond sigmoids: How to obtain well-calibrated probabilities from binary classifiers with beta calibration (2017, background): “A previous, shorter version of this paper appeared as (Kull, Silva Filho and Flach, 2017).”
A tutorial on calibration measurements and calibration models for clinical prediction models (2020, methodology): “39 When the logistic parametric assumptions are not met, properly calibrated estimates could suffer and become improperly calibrated.”
Beyond temperature scaling: Obtaining well-calibrated multiclass probabilities with Dirichlet calibration (2019, background): “Logit space Class probability space Binary classification Platt scaling [1] Beta calibration([2])”
Learning from positive and unlabeled data: a survey (2018, methodology): “Methods such as Platt scaling [82], isotonic regression [116] or beta calibration [50] can be used to calibrate classifiers that do not output well-calibrated probabilities.”
Multivariate Confidence Calibration for Object Detection (2020, methodology): “Logistic calibration is obtained using univariate normal distributions for g(s) [9].”
Deep generative models for reject inference in credit scoring (2019, methodology): “…which includes applications In addition to these two data sets, we have two small samples after September 2012 and December 2016 for Lending Club and Santander Bank respectively, which are used to produce well-calibrated estimates of class probabilities using the beta calibration approach [49].”
Aleatoric and epistemic uncertainty in machine learning: an introduction to concepts and methods (2019, methodology): “Examples of such methods include binning (Zadrozny and Elkan, 2001), isotonic regression (Zadrozny and Elkan, 2002), logistic scaling (Platt, 1999) and improvements thereof (Kull et al., 2017), as well as the use of Venn predictors (Johansson et al.”