An initialization method for the K-Means algorithm using neighborhood model.
Fuyuan CaoJiye LiangGuang JiangPublished in: Comput. Math. Appl. (2009)
Keyphrases
- k means
- cost function
- mathematical model
- clustering method
- em algorithm
- energy function
- expectation maximization
- high accuracy
- probabilistic model
- optimization model
- theoretical analysis
- recognition algorithm
- objective function
- input data
- hyper graph
- algorithm employs
- classification algorithm
- similarity measure
- dynamic programming
- parameter estimation
- estimation algorithm
- clustering algorithm
- prior information
- selection algorithm
- final result
- computational complexity
- improved algorithm
- classification method
- detection algorithm
- markov model
- support vector machine svm
- reconstruction method
- optimization method
- detection method
- preprocessing
- model free
- multiple models
- optimization algorithm
- tree structure
- parameter space
- hierarchical clustering
- computational cost
- bayesian framework
- segmentation method
- verification method
- sampling algorithm
- initial cluster centers
- data clustering
- learning algorithm
- initial guess
- spectral clustering
- filtering algorithm
- significant improvement
- kalman filter
- segmentation algorithm
- unsupervised clustering
- linear model
- neighborhood information
- simulated annealing
- hierarchical clustering algorithm
- matching algorithm
- fuzzy c means
- graph cuts
- levenberg marquardt
- closed form
- robust statistical
- convergence rate
- image segmentation