Embedded Feature Selection on Graph-Based Multi-View Clustering

Authors: Wenhui Zhao, Guangfei Li, Haizhou Yang, Quanxue Gao, Qianqian Wang

AAAI 2024 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Experimental results on five real-world datasets demonstrate that our proposed method outperforms state-of-the-art approaches.
Researcher Affiliation Academia Wenhui Zhao1, Guangfei Li1, Haizhou Yang1, Quanxue Gao1*, Qianqian Wang1 1 School of Telecommunication Engineering, Xidian University, Shaanxi 710071, China
Pseudocode Yes Algorithm 1: EFSGMC Input: Data matrices: {Xv}V v=1 RN dv, anchors number M, and cluster number K. Output: Cluster assignment matrix e G with K classes. 1: Initialize W = J = 0, Kv = Mv = 0, ρ0, ρ1, ρ2, pho ρ = 1.1, max ρi = 1010, αv = 1 V , γ, λ, β. 2: while not converg do 3: Update Qv by using (15) 4: Update Hv by using (30) 5: Update J by using (34) 6: Update Gv by solving (36) 7: Update Pv by solving (37) 8: Update αv by using (41) 9: Update Kv, Mv, W and ρi by using (42), (43), (44) and (45), respectively; 10: Directly achieve the K clusters based on the cluster assignment matrix e G=PV v=1 Gv αv ; 11: end while 12: return Clustering results.
Open Source Code No The paper does not provide an explicit statement about releasing source code for the described methodology or a link to a code repository.
Open Datasets Yes Datasets As shown in Table 1, MSRC (Winn and Jojic 2005), HW (Dua and Graff 2017), Mnist4 (Deng 2012), Cal101-20 (Fei-Fei, Fergus, and Perona 2007) and Reuters (Apt e, Damerau, and Weiss 1994) are selected.
Dataset Splits No The paper mentions hyperparameters are selected in a range to obtain optimal results, but it does not provide specific details on training, validation, or test dataset splits (percentages, counts, or explicit splitting methodology).
Hardware Specification No The paper does not provide specific hardware details (e.g., GPU/CPU models, processor types, memory amounts) used for running its experiments.
Software Dependencies No The paper does not provide specific ancillary software details, such as library or solver names with version numbers, needed to replicate the experiment.
Experiment Setup Yes Settings The hyperparameters are selected in the range of [0.0001, 0.001, 0.01, 0.1, 0.5, 1, 5, 10, 50, 100, 1000, 10000] to obtain the optimal results. Adjusting K within the range of 2-10, the optimal value of 5 was chosen.