A novel method for selecting initial centroids in K-means clustering algorithm.
S. PoomagalP. SaranyaS. KarthikPublished in: Int. J. Intell. Syst. Technol. Appl. (2016)
Keyphrases
- high precision
- computational complexity
- k means
- prior knowledge
- high accuracy
- selection algorithm
- support vector machine svm
- computational cost
- evaluation method
- fully automatic
- distance function
- synthetic data
- detection method
- main contribution
- edge detection
- experimental evaluation
- significant improvement
- pairwise
- dynamic programming
- input data
- evolutionary algorithm
- optimization algorithm
- preprocessing
- support vector
- matching algorithm
- objective function
- similarity measure
- image sequences
- clustering algorithm