Fast Multi-view Discrete Clustering with Anchor Graphs

Authors: Qianyao Qiang, Bin Zhang, Fei Wang, Feiping Nie9360-9367

AAAI 2021 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Extensive experiments on benchmark datasets demonstrate its efficiency and effectiveness. Comprehensive experiments are conducted on benchmark datasets to verify the proposed model.
Researcher Affiliation Academia Qianyao Qiang, 1 Bin Zhang, 1 Fei Wang, 2 Feiping Nie 3 1School of Software, Xi an Jiaotong University, Xi an 710049, China 2School of Electronics and Information Engineering, Xi an Jiaotong University, Xi an 710049, China 3School of Computer Science and Center for OPTical IMagery Analysis and Learning (OPTIMAL), Northwestern Polytechnical University, Xi an 710072, Shaanxi, P. R. China
Pseudocode Yes Algorithm 1 Algorithm to solve the problem in Eq. (17) Algorithm 2 Algorithm to solve the problem in Eq. (25)
Open Source Code No The paper does not provide any explicit statement or link for open-source code for the methodology described.
Open Datasets Yes The proposed model is evaluated on five widely used benchmark datasets with different scales and cluster numbers: BBCSport (Greene and Cunningham 2006), Web KB (Sun and Chao 2013), Reuters1, MNIST (Le Cun et al. 1998) and NUSWIDE (Chua et al. 2009).
Dataset Splits No All experiments are repeated tenfold and we report the average results and running time.
Hardware Specification Yes All the experiments are implemented on a Windows 10 desktop computer with a 3.6 GHz Intel Core i7-7700 CPU, 64 GB RAM and Matlab R2018b (64 bit).
Software Dependencies Yes All the experiments are implemented on a Windows 10 desktop computer with a 3.6 GHz Intel Core i7-7700 CPU, 64 GB RAM and Matlab R2018b (64 bit).
Experiment Setup Yes The anchors need to be sufficiently dense for effective adjacency relations, so we set anchor number m according to the scale in different datasets: m was set to 128 on BBCSports, Web KB and Reuters1, and 1024 on MIST, Reuters2 and NUS. The proposed model is not terminated until the difference of objective is less than 1e 10.