TY - GEN
T1 - Sparsity-constrained Graph Nonnegative Matrix Factorization for Clustering
AU - Chen, Keyi
AU - Che, Hangjun
AU - Yang, Xuanhao
AU - Leung, Man Fai
N1 - Funding Information:
The work is supported by National Natural Science Foundation of China(Grant No. 62003281) and the Fundamental Research Funds for the Central Universities (Grant No.SWU020006), and in part by Hong Kong Metropolitan University Research Grant (No. 2020/1.4).
Publisher Copyright:
© 2021 IEEE.
PY - 2021
Y1 - 2021
N2 - Graph nonnegative matrix factorization (GNMF) is superior for mining the intrinsic geometric structure embedded in high-dimensional data. As the sparsity of the factorized matrices is crucial for clustering, the l0 norm is commonly used in the formulated optimization problem to enforce the sparseness which makes the problem NP-hard and discontinuous. In this paper, the sparse graph nonnegative matrix factorization (SGNMF) is formulated as a global optimization problem by using the sum of inverted Gaussian functions to approximate the l0 norm, the multiplicative update rules are developed to solve the problem with guaranteed convergence. The superior performance of the proposed approach is substantiated by clustering tests on four public datasets.
AB - Graph nonnegative matrix factorization (GNMF) is superior for mining the intrinsic geometric structure embedded in high-dimensional data. As the sparsity of the factorized matrices is crucial for clustering, the l0 norm is commonly used in the formulated optimization problem to enforce the sparseness which makes the problem NP-hard and discontinuous. In this paper, the sparse graph nonnegative matrix factorization (SGNMF) is formulated as a global optimization problem by using the sum of inverted Gaussian functions to approximate the l0 norm, the multiplicative update rules are developed to solve the problem with guaranteed convergence. The superior performance of the proposed approach is substantiated by clustering tests on four public datasets.
KW - inverted Gaussian function
KW - multiplicative update rules
KW - Sparse graph nonnegative matrix
UR - http://www.scopus.com/inward/record.url?scp=85123843994&partnerID=8YFLogxK
U2 - 10.1109/ICICIP53388.2021.9642215
DO - 10.1109/ICICIP53388.2021.9642215
M3 - Conference contribution
AN - SCOPUS:85123843994
T3 - 11th International Conference on Intelligent Control and Information Processing, ICICIP 2021
SP - 292
EP - 299
BT - 11th International Conference on Intelligent Control and Information Processing, ICICIP 2021
T2 - 11th International Conference on Intelligent Control and Information Processing, ICICIP 2021
Y2 - 3 December 2021 through 7 December 2021
ER -