Adjusting the Outputs of a Classifier to New a Priori Probabilities: A Simple Procedure
Open accessThe DOI matches its Crossref record; 26 of the 26 references it cites were found in a registry or an index.
Abstract
It sometimes happens (for instance in case control studies) that a classifier is trained on a data set that does not reflect the true a priori probabilities of the target classes on real-world data. This may have a negative effect on the classification accuracy obtained on the real-world data set, especially when the classifier's decisions are based on the a posteriori probabilities of class membership. Indeed, in this case, the trained classifier provides estimates of the a posteriori probabilities that are not valid for this real-world data set (they rely on the a priori probabilities of the training set). Applying the classifier as is (without correcting its outputs with respect to these new conditions) on this new data set may thus be suboptimal. In this note, we present a simple iterative procedure for adjusting the outputs of the trained classifier with respect to these new a priori probabilities without having to refit the model, even when these probabilities are not known in advance. As a by-product, estimates of the new a priori probabilities are also obtained. This iterative algorithm is a straightforward instance of the expectation-maximization (EM) algorithm and is shown to maximize the likelihood of the new data. Thereafter, we discuss a statistical test that can be applied to decide if the a priori class probabilities have changed from the training set to the real-world data. The procedure is illustrated on different classification problems involving a multilayer neural network, and comparisons with a standard procedure for a priori probability estimation are provided. Our original method, based on the EM algorithm, is shown to be superior to the standard one for a priori probability estimation. Experimental results also indicate that the classifier with adjusted outputs always performs better than the original one in terms of classification accuracy, when the a priori probability conditions differ from the training set to the real-world data. The gain in classification accuracy can be significant.
Overview
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
In the citing paper
“However, if wellcalibrated fraud probabilities are needed, then the posterior probabilities can be adjusted (Saerens et al., 2002).”
From Data engineering for fraud detection (2021) · cites this paper as Saerens et al., 2002
Citations
Cites
Maximum Likelihood from Incomplete Data Via the EM Algorithm
Dempster, A. P., Laird, N. M., & Rubin, D. B. (1977) · Journal of the Royal Statistical Society Series B: Statistical Methodology · doi:10.1111/j.2517-6161.1977.tb01600.x · Cited by 50,444
Probability, Random Variables, and Stochastic Processes.
Abrahams, J., & Papoulis, A. (1984) · Journal of the American Statistical Association · doi:10.2307/2288754 · Cited by 16,226
UCI Repository of machine learning databases
Blake, C. L. (1998) · Medical Entomology and Zoology · Cited by 10,582
The EM Algorithm and Extensions
Kushary, D. (1998) · Technometrics · doi:10.1080/00401706.1998.10485534 · Cited by 5,417
An Introduction to the Theory of Statistics
(1911) · Nature · doi:10.1038/087345a0 · Cited by 4,394
Probability, Random Variables, and Stochastic Processes
Papoulis, A., & Hoffman, J. G. (1967) · Physics Today · doi:10.1063/1.3034123 · Cited by 4,107
Journal of the Royal Statistical Society (B)
Fildes, R. (1993) · International Journal of Forecasting · doi:10.1016/0169-2070(93)90088-5 · Cited by 3,437
Text classification from labeled and unlabeled documents using EM
Nigam, K., Mccallum, A. K., Thrun, S., & Mitchell, T. (2000) · Machine Learning · doi:10.1023/a:1007692713085 · Cited by 2,736 · cited 1×
“For instance, Nigam et al. (2000) use the EM algorithm to learn classiers that take advantage of both labeled and unlabeled data.”
MethodDiscriminant Analysis and Statistical Pattern Recognition.
Hawkins, D. M., & McLachlan, G. J. (1993) · Journal of the American Statistical Association · doi:10.2307/2290355 · Cited by 2,633
Mixture models : inference and applications to clustering
McLachlan, G. J., & Basford, K. E. (1988) · M. Dekker eBooks · Cited by 2,094
Cited by
The importance of intrinsic disorder for protein phosphorylation
Iakoucheva, L. M. (2004) · Nucleic Acids Research · doi:10.1093/nar/gkh253 · Cited by 1,429
“Predictors trained using a specific class distribution are known not to be optimal when applied to unlabeled data with different class priors ( 37 ).”
A unifying view on dataset shift in classification
Moreno-Torres, J. G., Raeder, T., Alaiz-Rodríguez, R., Chawla, N. V., & Herrera, F. (2012) · Pattern Recognition · doi:10.1016/j.patcog.2011.06.019 · Cited by 1,029
Learning when training data are costly: the effect of class distribution on tree induction
Weiss, G. M., & Provost, F. (2003) · Journal of Artificial Intelligence Research · doi:10.1613/jair.1199 · Cited by 921
Calibrating Probability with Undersampling for Unbalanced Classification
Pozzolo, A. D., Caelen, O., Johnson, R. A., & Bontempi, G. (2015) · 2015 IEEE Symposium Series on Computational Intelligence · doi:10.1109/ssci.2015.33 · Cited by 667
“[21] show that, given different priors between the training and testing sets, the posterior probability can be corrected with the following equation:”
BackgroundMethodResultInfluentialIdentification, analysis, and prediction of protein ubiquitination sites
Radivojac, P., Vacic, V., Haynes, C., Cocklin, R. R., Mohan, A., Heyen, J. W., Goebl, M. G., & Iakoucheva, L. M. (2010) · Proteins: Structure, Function, and Bioinformatics · doi:10.1002/prot.22555 · Cited by 583
Metadata
- Authors
- Saerens, M., Latinne, P., & Decaestecker, C.
- Affiliations
- Université Libre de Bruxelles (BE)
- Venue
- Neural Computation
- Publisher
- MIT Press
- Published
- 2002-01-01
- Volume
- 14
- Issue
- 1
- Pages
- 21-41
- ISSN
- 0899-7667 (ISSN-L)
- 1530-888X
- OpenAlex
- W2141253686
- Cited by
- 361
- References
- 26
- Citations by year
- 2026: 57 · 2025: 28 · 2024: 26 · 2023: 24 · 2022: 23 · 2021: 32 · 2020: 27 · 2019: 19 · 2018: 15 · 2017: 9 · 2016: 17 · 2015: 11 · 2014: 9 · 2013: 4 · 2012: 6
- FWCI
- 4.14
- Type
- journal article
- Language
- en
- Open access
- green
- dipot.ulb.ac.be
- Licence
- other-oa
- Locations
- Neural Computation (publishedVersion)
- PubMed (publishedVersion)
- Digital Access to Libraries (Université catholique de Louvain (UCL), l'Université de Namur (UNamur) and the Université Saint-Louis (USL-B)) (submittedVersion)
- Dépôt institutionnel de l'Université libre de Bruxelles (Université Libre de Bruxelles) (submittedVersion, other-oa)
- Topics
- Physical Sciences › Computer Science › Artificial Intelligence › Imbalanced Data Classification Techniques
- Physical Sciences › Computer Science › Artificial Intelligence › Machine Learning and Data Classification
- Physical Sciences › Computer Science › Artificial Intelligence › Data Stream Mining Techniques
- Keywords
- posterior probability, expectation-maximization algorithm, likelihood maximization, multilayer neural networks, case-control study, classification accuracy
- MeSH terms
- Classification, Probability, Probability Theory
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
How to cite
Cite CoChat for the checks on this page, and the paper for its own findings.
- This page
- CoChat. (2026). Citation check: Adjusting the Outputs of a Classifier to New a Priori Probabilities: A Simple Procedure. https://app.cochat.ai/papers/10.1162/089976602753284446
- The paper
- Saerens, M., Latinne, P., & Decaestecker, C. (2002). Adjusting the Outputs of a Classifier to New a Priori Probabilities: A Simple Procedure. Neural Computation. https://doi.org/10.1162/089976602753284446
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.