# Shift Happens: Adjusting Classifiers

arXiv: https://arxiv.org/abs/2111.02529 · HTML: https://mlsanity.com/paper/2111.02529/

## Key facts

- Cited by 3 papers in the map (data through 14 Oct 2026).
- Builds on 4 of its 20 references that are in the map.
- First posted on arXiv 3 Nov 2021.

## Builds on (most cited first, 4 of 4)

- [Domain-Adversarial Training of Neural Networks](https://mlsanity.com/paper/1505.07818/) (2015)
- [Convex optimization](https://mlsanity.com/paper/2106.01946/) (2021)
- [CVXPY: A Python-Embedded Modeling Language for Convex Optimization](https://mlsanity.com/paper/1603.00943/) (2016)
- [OpenML: networked science in machine learning](https://mlsanity.com/paper/1407.7722/) (2014)

## Built on by (most cited first, 3 of 3)

- [From Concept Drift to Model Degradation: An Overview on Performance-Aware Drift Detectors](https://mlsanity.com/paper/2203.11070/) (2022)
- [Improving the Reliability of Network Intrusion Detection Systems through Dataset Integration](https://mlsanity.com/paper/2112.02080/) (2021)
- [Prior-matched evaluation of operational Earth-observation classifiers: a three-number reporting method demonstrated on Sentinel-1 internal-wave detection](https://mlsanity.com/paper/2607.07146/) (2026)

## Abstract (arXiv)

Minimizing expected loss measured by a proper scoring rule, such as Brier score or log-loss (cross-entropy), is a common objective while training a probabilistic classifier. If the data have experienced dataset shift where the class distributions change post-training, then often the model's performance will decrease, over-estimating the probabilities of some classes while under-estimating the others on average. We propose unbounded and bounded general adjustment (UGA and BGA) methods that transform all predictions to (re-)equalize the average prediction and the class distribution. These methods act differently depending on which proper scoring rule is to be minimized, and we have a theoretical guarantee of reducing loss on test data, if the exact class distribution is known. We also demonstrate experimentally that, when in practice the class distribution is known only approximately, there is often still a reduction in loss depending on the amount of shift and the precision to which the class distribution is known.