We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:

Download:

Current browse context:

cs.LG

Change to browse by:

References & Citations

DBLP - CS Bibliography

Bookmark

(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo ScienceWISE logo

Computer Science > Machine Learning

Title: A Multi-Stage Automated Online Network Data Stream Analytics Framework for IIoT Systems

Abstract: Industry 5.0 aims at maximizing the collaboration between humans and machines. Machines are capable of automating repetitive jobs, while humans handle creative tasks. As a critical component of Industrial Internet of Things (IIoT) systems for service delivery, network data stream analytics often encounter concept drift issues due to dynamic IIoT environments, causing performance degradation and automation difficulties. In this paper, we propose a novel Multi-Stage Automated Network Analytics (MSANA) framework for concept drift adaptation in IIoT systems, consisting of dynamic data pre-processing, the proposed Drift-based Dynamic Feature Selection (DD-FS) method, dynamic model learning & selection, and the proposed Window-based Performance Weighted Probability Averaging Ensemble (W-PWPAE) model. It is a complete automated data stream analytics framework that enables automatic, effective, and efficient data analytics for IIoT systems in Industry 5.0. Experimental results on two public IoT datasets demonstrate that the proposed framework outperforms state-of-the-art methods for IIoT data stream analytics.
Comments: Published in IEEE Transactions on Industrial Informatics (Q1, IF: 11.648); Code is available at Github link: this https URL
Subjects: Machine Learning (cs.LG); Cryptography and Security (cs.CR); Networking and Internet Architecture (cs.NI)
MSC classes: 68T01, 60L10, 68W27
ACM classes: I.2.1; I.2.6; C.2.0
DOI: 10.1109/TII.2022.3212003
Cite as: arXiv:2210.01985 [cs.LG]
  (or arXiv:2210.01985v1 [cs.LG] for this version)

Submission history

From: Li Yang [view email]
[v1] Wed, 5 Oct 2022 02:18:36 GMT (4016kb,D)

Link back to: arXiv, form interface, contact.