Estimating regression errors without ground truth values
This addresses a practical challenge in regression analysis for applications where true outcomes are unknown, though it appears incremental as it builds on existing error estimation concepts.
The paper tackles the problem of estimating generalization error in regression when ground truth is unavailable, presenting a framework that performs robustly and is useful for detecting concept drift in real-world datasets.
Regression analysis is a standard supervised machine learning method used to model an outcome variable in terms of a set of predictor variables. In most real-world applications we do not know the true value of the outcome variable being predicted outside the training data, i.e., the ground truth is unknown. It is hence not straightforward to directly observe when the estimate from a model potentially is wrong, due to phenomena such as overfitting and concept drift. In this paper we present an efficient framework for estimating the generalization error of regression functions, applicable to any family of regression functions when the ground truth is unknown. We present a theoretical derivation of the framework and empirically evaluate its strengths and limitations. We find that it performs robustly and is useful for detecting concept drift in datasets in several real-world domains.