A Systematic Literature Review on Multi-label Data Stream Classification

TL;DR

This paper systematically reviews multi-label data stream classification methods, analyzing their approaches to challenges like concept drift, label emergence, and evaluation strategies, and identifies gaps for future research.

Contribution

It provides a comprehensive overview and hierarchical classification of recent methods, discussing their strategies, complexities, and resource use, and highlights research gaps.

Findings

Most methods address concept drift and label emergence.

Evaluation strategies vary widely across studies.

Identified key gaps and future directions in the field.

Abstract

Classification in the context of multi-label data streams represents a challenge that has attracted significant attention due to its high real-world applicability. However, this task faces problems inherent to dynamic environments, such as the continuous arrival of data at high speed and volume, changes in the data distribution (concept drift), the emergence of new labels (concept evolution), and the latency in the arrival of ground truth labels. This systematic literature review presents an in-depth analysis of multi-label data stream classification proposals. We characterize the latest methods in the literature, providing a comprehensive overview, building a thorough hierarchy, and discussing how the proposals approach each problem. Furthermore, we discuss the adopted evaluation strategies and analyze the methods' asymptotic complexity and resource consumption. Finally, we identify the main gaps and offer recommendations for future research directions in the field.