k-means++: few more steps yield constant approximation.
Davin ChooChristoph GrunauJulian PortmannVáclav RozhonPublished in: ICML (2020)
Keyphrases
- k means
- clustering algorithm
- approximation algorithms
- approximation error
- unsupervised clustering
- closed form
- approximation methods
- relative error
- spectral clustering
- data clustering
- approximation schemes
- special case
- real time
- approximation ratio
- sufficient statistics
- error tolerance
- database
- error bounds
- clustering method
- neural network
- data sets