Drift Estimation with Graphical Models

TL;DR

This paper introduces a graphical model-based approach to detect and quantify concept drift in datasets over time, independent of specific supervised models, demonstrated on real-world electric market data.

Contribution

It proposes a novel method that uses graphical model structure changes to assess concept drift without relying on the underlying supervised learning model.

Findings

Effective detection of concept drift through link creation and removal in graphical models

The method provides a stability metric for dataset evolution

Successful application to Australian Electric market data

Abstract

This paper deals with the issue of concept drift in supervised machine learn-ing. We make use of graphical models to elicit the visible structure of the dataand we infer from there changes in the hidden context. Differently from previous concept-drift detection methods, this application does not depend on the supervised machine learning model in use for a specific target variable, but it tries to assess the concept drift as independent characteristic of the evolution of a dataset. Specifically, we investigate how a graphical model evolves by looking at the creation of new links and the disappearing of existing ones in different time periods. The paper suggests a method that highlights the changes and eventually produce a metric to evaluate the stability over time. The paper evaluate the method with real world data on the Australian Electric market.