A Novel Method for Identifying Optimal Number of Clusters with Marginal Differential Entropy.
Bo ShuWei ChenZhendong NiuChangmin ZhangXiaotian JiangPublished in: WAIM Workshops (2013)
Keyphrases
- dynamic programming
- significant improvement
- synthetic data
- initial set
- hierarchical clustering
- preprocessing
- cost function
- high accuracy
- minimum error
- entropy measure
- closed form
- detection method
- neural network
- experimental evaluation
- computational complexity
- classification accuracy
- classification method
- data sets
- threshold values
- similarity measure
- exhaustive search
- bayesian networks
- high precision
- segmentation method
- pairwise
- clustering method
- mutual information
- computational cost
- support vector machine
- probability distribution