Extreme Gradient Boosted Multi-label Trees for Dynamic Classifier Chains

TL;DR

This paper introduces a fast, scalable multi-label classification method combining dynamic classifier chains with XGBoost, allowing for instance-specific label ordering and improved control over prediction dependencies, with reduced training costs.

Contribution

It presents a novel integration of dynamic classifier chains with XGBoost for multi-label classification, enabling instance-dependent label ordering and efficient training.

Findings

Effective on eleven datasets with improved control over label dependencies.

Reduces training costs by focusing on positive labels.

Demonstrates scalability and flexibility of the approach.

Abstract

Classifier chains is a key technique in multi-label classification, since it allows to consider label dependencies effectively. However, the classifiers are aligned according to a static order of the labels. In the concept of dynamic classifier chains (DCC) the label ordering is chosen for each prediction dynamically depending on the respective instance at hand. We combine this concept with the boosting of extreme gradient boosted trees (XGBoost), an effective and scalable state-of-the-art technique, and incorporate DCC in a fast multi-label extension of XGBoost which we make publicly available. As only positive labels have to be predicted and these are usually only few, the training costs can be further substantially reduced. Moreover, as experiments on eleven datasets show, the length of the chain allows for a more control over the usage of previous predictions and hence over the measure one want to optimize.