Enhash: A Fast Streaming Algorithm For Concept Drift Detection

TL;DR

Enhash is a fast, resource-efficient ensemble algorithm that detects various types of concept drift in data streams with competitive accuracy and speed.

Contribution

It introduces a novel projection hash-based ensemble method for rapid concept drift detection in streaming data.

Findings

Enhash performs comparably to existing methods in accuracy.

It requires less computational time.

It handles multiple drift types effectively.

Abstract

We propose Enhash, a fast ensemble learner that detects \textit{concept drift} in a data stream. A stream may consist of abrupt, gradual, virtual, or recurring events, or a mixture of various types of drift. Enhash employs projection hash to insert an incoming sample. We show empirically that the proposed method has competitive performance to existing ensemble learners in much lesser time. Also, Enhash has moderate resource requirements. Experiments relevant to performance comparison were performed on 6 artificial and 4 real data sets consisting of various types of drifts.