Research on self-training enhanced semi-supervised support vector machine framework

Weiquan Lu; Xuanzi Zhang; Lifeng Han; Jiarong Zhang; Yongheng Yang; Xintian Cheng; Yating Sun

doi:10.1117/12.3038653

22 July 2024 Research on self-training enhanced semi-supervised support vector machine framework

Weiquan Lu, Xuanzi Zhang, Lifeng Han, Jiarong Zhang, Yongheng Yang, Xintian Cheng, Yating Sun

Proceedings Volume 13222, International Conference on Signal Processing and Communication Security (ICSPCS 2024); 132221C (2024) https://doi.org/10.1117/12.3038653
Event: Third International Conference on Signal Processing and Communication Security (ICSPCS 2024), 2024, Chengdu, China

Abstract

This paper proposes a semi-supervised learning framework (SSL-SVM) based on Support Vector Machine (SVM) to address the complex features of modern data sets, such as high dimensionality and small sample sizes. The framework leverages unlabeled data through a self-training strategy, effectively addressing the issues of data scarcity and class imbalance, while improving the model's generalization ability, interpretability, and robustness. Experiments on liver cirrhosis datasets demonstrate that SSL-SVM is 14% more accurate than traditional SVM, providing a new strategy for data classification research.

(2024) Published by SPIE. Downloading of the abstract is permitted for personal use only.

Citation Download Citation

Weiquan Lu, Xuanzi Zhang, Lifeng Han, Jiarong Zhang, Yongheng Yang, Xintian Cheng, and Yating Sun "Research on self-training enhanced semi-supervised support vector machine framework", Proc. SPIE 13222, International Conference on Signal Processing and Communication Security (ICSPCS 2024), 132221C (22 July 2024); https://doi.org/10.1117/12.3038653

ACCESS THE FULL ARTICLE

INSTITUTIONAL
Select your institution to access the SPIE Digital Library.

SELECT YOUR INSTITUTION

PERSONAL
Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.

PERSONAL SIGN IN

No SPIE Account? Create one

PURCHASE THIS CONTENT

SUBSCRIBE TO DIGITAL LIBRARY

50 downloads per 1-year subscription

Members: $195

Non-members: $335 ADD TO CART

25 downloads per 1 - year subscription

Members: $145

Non-members: $250 ADD TO CART

PURCHASE SINGLE ARTICLE

Includes PDF, HTML & Video, when available