A modification of the k-means method for quasi-unsupervised learning.
David Rebollo-MonederoMarc SoléJordi NinJordi FornéPublished in: Knowl. Based Syst. (2013)
Keyphrases
- unsupervised learning
- k means
- synthetic data
- detection method
- clustering method
- cost function
- high precision
- preprocessing
- pairwise
- significant improvement
- experimental evaluation
- dynamic programming
- classification method
- high accuracy
- classification accuracy
- objective function
- computational cost
- neural network
- prior knowledge
- multiresolution
- computationally efficient
- segmentation algorithm
- pattern recognition
- multiscale
- segmentation method
- clustering algorithm
- data mining