An empirical comparison of four initialization methods for the K-Means algorithm.
José M. PeñaJosé Antonio LozanoPedro LarrañagaPublished in: Pattern Recognit. Lett. (1999)
Keyphrases
- k means
- preprocessing
- clustering algorithm
- computational cost
- learning algorithm
- expectation maximization
- benchmark data sets
- spectral clustering
- significant improvement
- clustering method
- recently published
- synthetic and real datasets
- synthetic and real images
- dynamic programming
- data clustering
- clustering quality
- computational complexity
- experimental evaluation
- simulated annealing
- heuristic methods
- worst case
- segmentation algorithm
- optimization algorithm
- parameter settings
- noisy data
- exhaustive search
- data sets
- recognition algorithm
- theoretical guarantees
- hierarchical clustering
- matching algorithm
- graph cuts
- maximum likelihood
- input data
- support vector machine
- probabilistic model
- search algorithm
- decision trees