Bias Correction in Machine Learning-based Classification of Rare Events

TL;DR

This paper presents a machine learning approach for classifying rare online platforms from web texts, focusing on reducing false positives and bias in probability estimates to improve detection accuracy.

Contribution

It introduces a bias correction method using calibrated probabilities and ensembles specifically for rare event classification in natural language processing.

Findings

Significant reduction in false positives.

Improved bias in probability estimates.

Enhanced detection accuracy for rare online platforms.

Abstract

Online platform businesses can be identified by using web-scraped texts. This is a classification problem that combines elements of natural language processing and rare event detection. Because online platforms are rare, accurately identifying them with Machine Learning algorithms is challenging. Here, we describe the development of a Machine Learning-based text classification approach that reduces the number of false positives as much as possible. It greatly reduces the bias in the estimates obtained by using calibrated probabilities and ensembles.