Spectral clustering via ensemble deep autoencoder learning (SC-EDAE). (December 2020)

• Record Type: Journal Article
• Authors: Affeldt, Séverine; Labiod, Lazhar; Nadif, Mohamed
• Is Part Of: Pattern recognition. Volume 108(2020:Dec.)
• Publication Date: 2020-12
• Subjects: Spectral clustering -- Unsupervised ensemble learning -- Autoencoder,; Pattern perception -- Periodicals; Perception des structures -- Périodiques; Patroonherkenning; 006.4
• DOI: 10.1016/j.patcog.2020.107522
• Languages: English
• ISSNs: 0031-3203

Abstract

Highlights: We propose a robust and efficient deep clustering approach with no pre-training. We combine spectral clustering, deep embeddings and ensemble paradigm strengths. Our original clustering method inherits the low complexity of landmarks strategy. The effectiveness is shown through extensive experiments on real-world datasets. Graphical abstract: Abstract: Several works have studied clustering strategies that combine classical clustering algorithms and deep learning methods. These strategies generally improve clustering performance, however deep autoencoder setting issues impede the robustness of these approaches. To alleviate the impact of hyperparameters setting, we propose a model which combines spectral clustering and deep autoencoder strengths in an ensemble framework. Our proposal does not require any pretraining and includes the three following steps: generating various deep embeddings from the original data, constructing a sparse and low-dimensional ensemble affinity matrix based on anchors strategy and applying spectral clustering to obtain the common space shared by multiple deep representations. While the anchors strategy ensures an efficient merging of the encodings, the fusion of various deep representations enables to mitigate the deep networks setting issues. Experiments on various benchmark datasets demonstrate the potential and robustness of our approach compared to state-of-the-art deep clustering methods.