Litcius/Paper detail

Unbalanced Incomplete Multi-View Clustering Via the Scheme of View Evolution: Weak Views are Meat; Strong Views Do Eat

Xiang Fang, Yuchong Hu, Pan Zhou, Dapeng Oliver Wu

2021IEEE Transactions on Emerging Topics in Computational Intelligence65 citationsDOIOpen Access PDF

Abstract

Incomplete multi-view clustering is an important technique to deal with real-world incomplete multi-view data. Previous methods assume that all views have the same incompleteness, i.e., balanced incompleteness. However, different views often have distinct incompleteness, i.e., unbalanced incompleteness, which results in <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">strong views</i> (low-incompleteness views) and <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">weak views</i> (high-incompleteness views). The unbalanced incompleteness prevents us from directly using previous methods. In this paper, inspired by the effective biological evolution theory, we design the novel scheme of <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">view evolution</i> to cluster strong and weak views. Moreover, we propose an Unbalanced Incomplete Multi-view Clustering method (UIMC), which is the <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">first</i> effective method based on view evolution for unbalanced incomplete multi-view clustering. Compared with previous methods, UIMC has two unique advantages: 1) it proposes weighted multi-view subspace clustering to integrate unbalanced incomplete views, which effectively solves the unbalanced incomplete multi-view clustering problem; 2) it designs the low-rank representation to recover the data, which diminishes the impact of the incompleteness and noises. Extensive experimental results demonstrate that UIMC improves the clustering performance by up to 40% on three evaluation metrics over other state-of-the-art methods. We provide codes for all of our experiments in <uri xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">https://github.com/ZeusDavide/TETCI_UIMC</uri> .

Topics & Concepts

Cluster analysisRepresentation (politics)Computer scienceScheme (mathematics)Constrained clusteringData miningComplete informationCluster (spacecraft)MathematicsFuzzy clusteringCorrelation clusteringAlgorithmMathematical optimizationArtificial intelligenceSubspace topologyCURE data clustering algorithmSingle-linkage clusteringMachine learningFace and Expression RecognitionAdvanced Clustering Algorithms ResearchImbalanced Data Classification Techniques