Self-weighted Multiview Clustering with Multiple Graphs

Authors: Feiping Nie, Jing Li, Xuelong Li

IJCAI 2017 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Evaluations on two synthetic datasets indicate the effectiveness of our methods. Compared with several representative graphbased multiview clustering approaches on four realworld datasets, the proposed methods achieve the better performances and our new clustering method is more practical to use.
Researcher Affiliation Academia 1School of Computer Science and Center for OPTical IMagery Analysis and Learning (OPTIMAL), Northwestern Polytechnical University, Xi an 710072, P. R. China 2Xi an Institute of Optics and Precision Mechanics, Chinese Academy of Sciences, Xi an 710119, P. R. China
Pseudocode Yes Algorithm 1 The algorithm of Self-weighted Multiview Clustering (Sw MC) in Eq. (4), Algorithm 2 The algorithm of Parameter-weighted Multiview Clustering (Pw MC) in Eq. (3)
Open Source Code No The paper does not provide any statement or link indicating the availability of open-source code for the described methodology.
Open Datasets Yes MSRCv1 [Winn and Jojic, 2005], Caltech101 [Fei-Fei et al., 2007] (we use two regular subsets Caltech101-7 and Caltech101-20), Handwritten numerals (Digits) [Asuncion and Newman, 2007]
Dataset Splits No The paper references the use of datasets for evaluation but does not specify training, validation, or test dataset splits (e.g., percentages or sample counts).
Hardware Specification No The paper does not provide any specific details about the hardware used for running the experiments.
Software Dependencies No The paper does not specify any software dependencies with version numbers.
Experiment Setup Yes In this paper, we follow CLR method to construct a graph for each view as the initialized input graph A(v). Different from some regular kernel-based methods [Zelnik-Manor and Perona, 2005; Cai et al., 2005], this graph construction approach only needs to set one parameter k which represents the number neighbors. For all the compared methods, we validate their performances by fixing k as 10. [...] for all the methods involving K-means, we repeat them for 50 times and report the averaged result. As for our methods, we only run once. [...] Typically, γ in Pw MC is searched in logarithm form (log10γ from 0 to 4 with step size 0.5).