Presentation Details
  1. Home
  2. Presentation List
  3. Presentation Details
Research on Fault Diagnosis of Bearing Sample Imbalance Based on HCAB-SMOTE
ID：153 View Protection：ATTENDEE Updated Time：2025-11-10 16:14:40 Hits：133 Poster Presentation

Start Time：Pending（Asia/Shanghai）

Duration：Pending

Session：No Session »

No files

Abstract
To addresses the issue of scarce samples and sample imbalance in marine bearing fault diagnosis, which leads to low diagnostic efficiency, and proposes a fault diagnosis method based on HCAB-SMOTE (Hybrid Clustering Boundary Synthetic Minority Over-sampling Technique). The method first extracts impact-related features from the original vibration signals; then, it utilizes HCAB-SMOTE to intelligently over-sample minority class fault samples, effectively alleviating the classifier bias problem caused by sample imbalance. Experiments were conducted using a simulated experimental dataset to compare the performance of four sampling methods, original data, SMOTE, Borderline-SMOTE, and HCAB-SMOTE, across SVM. The results indicate that in the SVM classifier, HCAB-SMOTE outperforms traditional oversampling methods in multiple key performance metrics, particularly in its ability to identify minority classes, proving its effectiveness and superiority in ship bearing fault diagnosis.
Keywords
Marine bearings; HCAB-SMOTE; sample imbalance; fault diagnosis;
Speaker
Yan Zhijia
postgraduate Guangdong Ocean university

Submission Author
Yan Zhijia Guangdong Ocean university
LIAO ZHIQIANG Guangdong Ocean University
Cai Renchao Guangdong Ocean University
Submit Comment
Verify Code Change Another
All Comments
Important Date

  • Conference Date

    Nov 21

    2025

    to

    Nov 23

    2025

  • Oct 20 2025

    Draft paper submission deadline

  • Dec 08 2025

    Registration deadline

Sponsored By
IEEE Instrumentation and Measurement Society
South China University of Technology
Organized By
South China University of Technology
Contact Information
Login to view full contact details
Previous Conferences