Progressive graph-based subspace transductive learning for semi-supervised classification

• Published in: IET image processing Vol. 13; no. 14; pp. 2753 - 2762
• Main Authors: Chen, Long, Zhong, Zhi
• Format: Journal Article
• Language: English
• Published: The Institution of Engineering and Technology 12.12.2019
• Subjects: efficient semisupervised learning technique; feature affinity matrix; feature domain; feature information; feature relationships; feature‐to‐label alignment; fixed subject‐wise graph; graph theory; high‐dimensional feature space; iterative methods; iterative optimisation strategy; label domain; label information; learning (artificial intelligence); matrix algebra; noise points; optimisation; pattern classification; PGSTL; progressive graph‐based subspace transductive learning; representative relation matrix; semisupervised classification; Special Section: Adversarial Learning in Image Processing; sufficient labelled samples; label domain; pattern classification; iterative methods; sufficient labelled samples; representative relation matrix; graph theory; high-dimensional feature space; feature information; semisupervised classification; feature affinity matrix; matrix algebra; noise points; progressive graph-based subspace transductive learning; feature relationships; optimisation; feature domain; efficient semisupervised learning technique; fixed subject-wise graph; feature-to-label alignment; learning (artificial intelligence); PGSTL; iterative optimisation strategy; label information
• ISSN: 1751-9659; 1751-9667
• DOI: 10.1049/iet-ipr.2018.6363

Graph-based transductive learning (GTL) is the efficient semi-supervised learning technique which is always employed in that sufficient labeled samples can not be obtained. Conventional GTL methods generally construct a inaccurate graph in feature domain and they are not able to align feature information with label information. To address these issues, we propose an approach called Progressive Graph-based subspace transductive learning (PGSTL) in this paper. PGSTL gradually find the intrinsic relationship between samples that more accurately aligns feature with label. Meanwhile, PGSTL develops a feature affinity matrix in the subspace of original high-dimensional feature space, which effectively reduce the interference of noise points. And then, the representative relation matrix and the feature affinity matrix are optimized by iterative optimization strategy and finally aligned. In this way, PGSTL can not only effectively reduce the interference of noisy points, but also comprehensively consider the information in the feature and label domain of data. Extensive experimental results on various benchmark datasets demonstrate that the PGSTL achieves the best performance compared to some state-of-the-art semi-supervised learning methods.