Lowering the complexity of k-means clustering by BFS-dijkstra method for graph computing.
Anna ZhangJun YaoYasuhiko NakashimaPublished in: COOL Chips (2015)
Keyphrases
- significant improvement
- graph matching
- synthetic data
- dynamic programming
- computational cost
- high accuracy
- cost function
- computational complexity
- experimental evaluation
- objective function
- pairwise
- segmentation method
- prior knowledge
- support vector machine
- support vector machine svm
- clustering method
- preprocessing
- high precision