Knowledge-enhanced Incomplete Multi-view Clustering on Relational Knowledge

Incomplete multi-view clustering (IMVC) methods have drawn increasingly attention these years, as their better potential to the complex real-world scenario. The key of the models in such tasks is how to perform high-quality imputation. Traditional intra-view and inter-view imputation strategies rely mainly on observed features. Such strategies may introduce semantic noise and biased estimations, making it difficult to recover the true semantics of missing views. Meanwhile, they are hard to directly apply to relational knowledge MVC, as fail to utilize the knowledge sufficiently. To this end, we propose a simple yet effective knowledge-enhanced anchor-based IMVC framework with two specific knowledge enhancement mechanisms, i.e., one for anchors, and the other for anchor graph. With the structural knowledge enhancement, the model can be gifted with better expressive ability and get good results. Extensive experiments compared with 10 models have proven the promising capacity of our method from 5 aspects, i.e., superiority, stability, sensitivity, complexity, and visualization. Notably, as the first work on IMVC on relational knowledge, 45 incomplete multi-view datasets are constructed with missing ratios ranging from 10% to 90%, which also facilitate the future research.

Authors

Institutions

Publication Details

Journal
IEEE Transactions on Pattern Analysis and Machine Intelligence
Published
2026-11-18
DOI
https://doi.org/10.1109/tpami.2026.3754812
Primary Topic
Advanced Clustering Algorithms Research
Type
article
Field-Weighted Citation Impact
0.00
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
article

Knowledge-enhanced Incomplete Multi-view Clustering on Relational Knowledge

Ke Liang, Weixuan Liang, Xinhang Wan, Miaomiao Li et al.
1 links
IEEE Transactions on Pattern Analysis and Machine Intelligence
Advanced Clustering Algorithms Research
article

Knowledge-enhanced Incomplete Multi-view Clustering on Relational Knowledge

Ke Liang, Weixuan Liang, Xinhang Wan, Miaomiao Li, Xinwang Liu, Tianrui Liu, Mengjiao Zhang, Hao Li
article en

Abstract

Incomplete multi-view clustering (IMVC) methods have drawn increasingly attention these years, as their better potential to the complex real-world scenario. The key of the models in such tasks is how to perform high-quality imputation. Traditional intra-view and inter-view imputation strategies rely mainly on observed features. Such strategies may introduce semantic noise and biased estimations, making it difficult to recover the true semantics of missing views. Meanwhile, they are hard to directly apply to relational knowledge MVC, as fail to utilize the knowledge sufficiently. To this end, we propose a simple yet effective knowledge-enhanced anchor-based IMVC framework with two specific knowledge enhancement mechanisms, i.e., one for anchors, and the other for anchor graph. With the structural knowledge enhancement, the model can be gifted with better expressive ability and get good results. Extensive experiments compared with 10 models have proven the promising capacity of our method from 5 aspects, i.e., superiority, stability, sensitivity, complexity, and visualization. Notably, as the first work on IMVC on relational knowledge, 45 incomplete multi-view datasets are constructed with missing ratios ranging from 10% to 90%, which also facilitate the future research.

IEEE Transactions on Pattern Analysis and Machine IntelligenceVol. PP
National University of Defense Technology (CN), Changsha University (CN)
Advanced Clustering Algorithms Research
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon