Presents FS-MWK++ and scalable SFS-MWK++ algorithms for unsupervised feature selection via stability of feature weights in Minkowski weighted k-means across a range of exponents, supported by theoretical analysis under noise and cluster assumptions.
Some methods for classification and analysis of multi- variate observations
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2025 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Scalable unsupervised feature selection via weight stability
Presents FS-MWK++ and scalable SFS-MWK++ algorithms for unsupervised feature selection via stability of feature weights in Minkowski weighted k-means across a range of exponents, supported by theoretical analysis under noise and cluster assumptions.