Unsupervised Feature Selection with Binary Hashing
Data
- single-view datasets:single-label and multi-label datasets
- multi-view datasets
UMFS Unsupervised Multi-label Feature Selection
MUMFS Multo-view extension: Multi-view Unsupervised Multi-label Feature Selection
Evaluation
- Clustering task: perform k-means
- Classification task: perform SVM (download libSVM package) and ML-KNN (use MLKNN package)
Evaluation Metrics
clustering: accuracy(ACC) and Normalized Mutual Information(NMI)(run 'ClusteringMeasure.m')
classification: accuracy; multi-label classification: One-Error and Average Precision.
@article{DBLP:journals/tip/ShiZLZC23,
author = {Dan Shi and
Lei Zhu and
Jingjing Li and
Zheng Zhang and
Xiaojun Chang},
title = {Unsupervised Adaptive Feature Selection With Binary Hashing},
journal = {{IEEE} Trans. Image Process.},
volume = {32},
pages = {838--853},
year = {2023}
}