TY - GEN
T1 - Revisiting Fast Spectral Clustering with Anchor Graph
AU - Wang, Cheng Long
AU - Nie, Feiping
AU - Wang, Rong
AU - Li, Xuelong
N1 - Publisher Copyright:
© 2020 IEEE.
PY - 2020/5
Y1 - 2020/5
N2 - Many anchor-graph-based spectral clustering methods have been proposed to accelerate spectral clustering for large scale problems. In this paper, we revisit the popular large-scale spectral clustering method based on the anchor graph which is equivalent to the spectral decomposition on a similar matrix obtained using a second-order transition probability. However, due to the special structure of the bipartite graph, there is no stable distribution of the random walk process. The even-order transition probabilities may only a side view of the bipartite structure, resulting in breaking the independence of data points and leading to undesired artifacts for boundary samples. Therefore, we propose a Fast Spectral Clustering based on the Random Walk Laplacian (FRWL) method. The random walk Laplacian balances explicitly the popularity of anchors and the independence of data points, which keeps the structure of boundary samples. The experimental results demonstrate the efficiency and effectiveness of our method.
AB - Many anchor-graph-based spectral clustering methods have been proposed to accelerate spectral clustering for large scale problems. In this paper, we revisit the popular large-scale spectral clustering method based on the anchor graph which is equivalent to the spectral decomposition on a similar matrix obtained using a second-order transition probability. However, due to the special structure of the bipartite graph, there is no stable distribution of the random walk process. The even-order transition probabilities may only a side view of the bipartite structure, resulting in breaking the independence of data points and leading to undesired artifacts for boundary samples. Therefore, we propose a Fast Spectral Clustering based on the Random Walk Laplacian (FRWL) method. The random walk Laplacian balances explicitly the popularity of anchors and the independence of data points, which keeps the structure of boundary samples. The experimental results demonstrate the efficiency and effectiveness of our method.
KW - Spectral Clustering
KW - anchor-based graph
KW - large scale clustering
KW - random walk Laplacian
UR - http://www.scopus.com/inward/record.url?scp=85089221777&partnerID=8YFLogxK
U2 - 10.1109/ICASSP40776.2020.9053271
DO - 10.1109/ICASSP40776.2020.9053271
M3 - 会议稿件
AN - SCOPUS:85089221777
T3 - ICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings
SP - 3902
EP - 3906
BT - 2020 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2020 - Proceedings
PB - Institute of Electrical and Electronics Engineers Inc.
T2 - 2020 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2020
Y2 - 4 May 2020 through 8 May 2020
ER -