Benchmarking Concept Drift Detectors for Online Machine Learning
Date
2022
Journal Title
Journal ISSN
Volume Title
Publisher
Springer, Cham
Abstract
Concept drift detection is an essential step to maintain the
accuracy of online machine learning. The main task is to detect changes
in data distribution that might cause changes in the decision bound aries for a classification algorithm. Upon drift detection, the classifica tion algorithm may reset its model or concurrently grow a new learning
model. Over the past fifteen years, several drift detection methods have
been proposed. Most of these methods have been implemented within
the Massive Online Analysis (MOA). Moreover, a couple of studies have
compared the drift detectors. However, such studies have merely focused
on comparing the detection accuracy. Moreover, most of these studies are
focused on synthetic data sets only. Additionally, these studies do not
consider drift detectors not integrated into MOA. Furthermore, None of
the studies have considered other metrics like resource consumption and
runtime characteristics. These metrics are of utmost importance from an
operational point of view.
In this paper, we fill this gap. Namely, this paper evaluates the perfor mance of sixteen different drift detection methods using three different
metrics: accuracy, runtime, and memory usage. To guarantee a fair com parison, MOA is used. Fourteen algorithms are implemented in MOA.
We integrate two new algorithms (ADWIN++ and SDDM) into MOA.
Description
Keywords
Online machine learning · Concept drifts · Benchmarking
Citation
Mahgoub, M., Moharram, H., Elkafrawy, P., Awad, A. (2023). Benchmarking Concept Drift Detectors for Online Machine Learning. In: Fournier-Viger, P., Hassan, A., Bellatreche, L. (eds) Model and Data Engineering. MEDI 2022. Lecture Notes in Computer Science, vol 13761. Springer, Cham.