MvWECM: Multi-view Weighted Evidential C-Means clustering

Preparation

Dependency

R == 4.4.1

(Rstudio will prompt you to install the missing packages.)

• aricode

• R.utils

• R.matlab

• evclust (optional)

Data

INPUT: views x n x d list

Datasets can be found at Link1, Link2 and Link3.

Usage

Just main.R. (Change your path and inputting datasets)

Parameters

Our platform is i7-13700 + 24G (Mac & Win). We recommend you set type = "pairs" on larger datasets for faster computation.

datasets	alpha	beta	delta	eta	preprocessing
Prok	1	1~403	2	1	normalize
Webkb	1	3~340	13~15	1	0-1
IS	1	241, 259	2~3	1	normalize
Caltech07	1	277	4	1	0-1
3sources	1	1	9~20	1	0-1
Reuters-1500	1	29	3	1	0-1
Reuters-18758	1.25	58	2	1	normalize

Remark:

If methods do not output credal partition, please change if_credal_id.

Citation

If you find MvWECM useful in your research, please consider citing:

BibTeX

@article{zhou2025,
  title        = {MvWECM: Multi-View Weighted Evidential C-Means Clustering},
  author       = {Kuang Zhou and Yuchen Zhu and Mei Guo and Ming Jiang},
  journal      = {Pattern Recognition},
  volume       = {159},
  pages        = {111108},
  year         = {2024},
  issn         = {0031-3203},
  publisher    = {Elsevier}
}