Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

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

Li Yang, Abdallah Shami

TL;DR

MSANA addresses concept drift in IIoT data streams under Industry 5.0 by delivering a four-stage automated analytics pipeline that combines dynamic data pre-processing, drift-aware feature selection, online base learners with dynamic selection, and a window-based ensemble (W-PWPAE). It introduces Drift-based Dynamic Feature Selection (DD-FS) and a Window-based Performance Weighted Probability Averaging Ensemble, and validates the framework on IoTID20 and CICIDS2017 datasets, achieving high accuracy (up to 99.32%) with low latency (as low as ~3.5 ms). The results demonstrate superior performance and efficiency compared with state-of-the-art online drift-adaptive methods, supporting real-time IIoT anomaly detection and security analytics. The framework is designed for deployment on edge-cloud architectures, enabling scalable, automated analytics in dynamic IIoT environments for Industry 5.0 applications.

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.

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

TL;DR

MSANA addresses concept drift in IIoT data streams under Industry 5.0 by delivering a four-stage automated analytics pipeline that combines dynamic data pre-processing, drift-aware feature selection, online base learners with dynamic selection, and a window-based ensemble (W-PWPAE). It introduces Drift-based Dynamic Feature Selection (DD-FS) and a Window-based Performance Weighted Probability Averaging Ensemble, and validates the framework on IoTID20 and CICIDS2017 datasets, achieving high accuracy (up to 99.32%) with low latency (as low as ~3.5 ms). The results demonstrate superior performance and efficiency compared with state-of-the-art online drift-adaptive methods, supporting real-time IIoT anomaly detection and security analytics. The framework is designed for deployment on edge-cloud architectures, enabling scalable, automated analytics in dynamic IIoT environments for Industry 5.0 applications.

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.
Paper Structure (18 sections, 10 equations, 6 figures, 2 tables, 2 algorithms)

This paper contains 18 sections, 10 equations, 6 figures, 2 tables, 2 algorithms.

Table of Contents

  1. Introduction
  2. Related Work
  3. Concept Drift Detection
  4. Concept Drift Adaptation
  5. Proposed Framework
  6. System Overview
  7. Dynamic Data Pre-Processing
  8. Data Balancing
  9. Data Normalization
  10. Drift-based Dynamic Feature Selection
  11. Model Learning
  12. Base Model Learning and Dynamic Selection
  13. Online Ensemble Model Development
  14. Performance Evaluation
  15. Experimental Setup
  16. ...and 3 more sections

Figures (6)

  • Figure 1: The framework of the proposed MSANA system.
  • Figure 2: Accuracy comparison of state-of-the-art drift adaptation methods on the CICIDS2017 dataset.
  • Figure 3: Accuracy comparison of state-of-the-art drift adaptation methods on the IoTID20 dataset.
  • Figure 4: Probability Density Function (PDF) of the inference time of the proposed framework on the CICIDS2017 dataset.
  • Figure 5: Probability Density Function (PDF) of the inference time of the proposed framework on the IoTID20 dataset.
  • ...and 1 more figures