Home > Research > Publications & Outputs > A Self-Training Hierarchical Prototype-Based Ap...

Electronic data

  • STHP_revison_v1

    Rights statement: This is the author’s version of a work that was accepted for publication in Information Science. Changes resulting from the publishing process, such as peer review, editing, corrections, structural formatting, and other quality control mechanisms may not be reflected in this document. Changes may have been made to this work since it was submitted for publication. A definitive version was subsequently published in Information Science, 535, 2020 DOI: 10.1016/j.ins.2020.05.018

    Accepted author manuscript, 1.55 MB, PDF document

    Embargo ends: 13/05/21

    Available under license: CC BY-NC-ND: Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License

Links

Text available via DOI:

View graph of relations

A Self-Training Hierarchical Prototype-Based Approach for Semi-Supervised Classification

Research output: Contribution to journalJournal articlepeer-review

Published
<mark>Journal publication date</mark>1/10/2020
<mark>Journal</mark>Information Sciences
Volume535
Number of pages21
Pages (from-to)204-224
Publication StatusPublished
Early online date13/05/20
<mark>Original language</mark>English

Abstract

This paper introduces a novel self-training hierarchical prototype-based approach for semi-supervised classification. The proposed approach firstly identifies meaningful prototypes from labelled samples at multiple levels of granularity and, then, self-organizes a highly transparent, multi-layered recognition model by arranging them in a form of pyramidal hierarchies. After this, the learning model continues to self-evolve its structure and self-expand its knowledge base to incorporate new patterns recognized from unlabelled samples by exploiting the pseudo-label technique. Thanks to its prototype-based nature, the overall computational process of the proposed approach is highly explainable and traceable. Experimental studies with various benchmark image recognition problems demonstrate the state-of-the-art performance of the proposed approach, showing its strong capability to mine key information from unlabelled data for classification.

Bibliographic note

This is the author’s version of a work that was accepted for publication in Information Science. Changes resulting from the publishing process, such as peer review, editing, corrections, structural formatting, and other quality control mechanisms may not be reflected in this document. Changes may have been made to this work since it was submitted for publication. A definitive version was subsequently published in Information Science, 535, 2020 DOI: 10.1016/j.ins.2020.05.018