Reliable Multi-View Clustering

Authors: Hong Tao, Chenping Hou, Xinwang Liu, Tongliang Liu, Dongyun Yi, Jubo Zhu

AAAI 2018 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Experimental results on a number of data sets demonstrate that the proposed method can effectively improve the reliability of multi-view clustering.
Researcher Affiliation Academia Hong Tao,1 Chenping Hou,1 Xinwang Liu,2 Dongyun Yi,1 Jubo Zhu1 1College of Science, National University of Defense Technology, Changsha, 410073, Hunan, China. 2School of Computer, National University of Defense Technology, Changsha, 410073, Hunan, China.
Pseudocode Yes Algorithm 1 summarizes the pseudocode of the proposed method.
Open Source Code No The paper does not provide a statement or link indicating that the source code for the proposed RMVC method is publicly available.
Open Datasets Yes The six text data sets are 3sources3 (3sou), BBC3/4views (BBC3/4) and BBCSport2/3/4views4 (BSpt2/3/4)., with footnotes like 3http://mlg.ucd.ie/datasets/3sources.html and 5http://www.vision.caltech.edu/Image Datasets/Caltech101/ providing access.
Dataset Splits No The paper does not provide specific details on training/validation/test dataset splits, such as percentages, sample counts, or explicit references to predefined splits used for validation.
Hardware Specification Yes The experiments are conducted on a work station with 12 cores (2.10 GHz for each) and 96.0 GB RAM memory.
Software Dependencies Yes All algorithms are tested with MATLAB R2013b, and our method also use MOSEK 7.1.
Experiment Setup Yes Two points are connected if at least one of them is among the k nearest neighbors of the other in the Euclidean distance and k is set to be 9 empirically. The edge weight is calculated using Gaussian Kernel, where the bandwidth parameter is set as the mean squared Euclidean distance between sample pairs. For Co Reg SC, Co Train SC and MMSC, their trade-off parameters are selected from {0.01, 0.1, 1, 10, 100}, and the best results are reported.