Sampling Bias Correction for Supervised Machine Learning: A Bayesian Inference Approach with Practical Applications

TL;DR

This paper introduces a Bayesian inference method to correct sampling bias in supervised learning, enabling models to better represent the original data distribution across various applications.

Contribution

It proposes a Bayesian approach to adjust for known sampling bias, applicable to multiple models and real-world scenarios.

Findings

Effective bias correction demonstrated on logistic regression.

Applicable to diverse fields like medicine, image recognition, and marketing.

Enhances inference accuracy from biased datasets.

Abstract

Given a supervised machine learning problem where the training set has been subject to a known sampling bias, how can a model be trained to fit the original dataset? We achieve this through the Bayesian inference framework by altering the posterior distribution to account for the sampling function. We then apply this solution to binary logistic regression, and discuss scenarios where a dataset might be subject to intentional sample bias such as label imbalance. This technique is widely applicable for statistical inference on big data, from the medical sciences to image recognition to marketing. Familiarity with it will give the practitioner tools to improve their inference pipeline from data collection to model selection.