WSD-algorithm based on new method of vector-word contexts proximity calculation via epsilon-filtration.
Alexander KirillovNatalia KrizhanovskyAndrew KrizhanovskyPublished in: CoRR (2018)
Keyphrases
- computational cost
- dynamic programming
- high accuracy
- cost function
- detection algorithm
- optimization algorithm
- preprocessing
- significant improvement
- improved algorithm
- k means
- detection method
- recognition algorithm
- clustering method
- optimization method
- computationally efficient
- computational complexity
- theoretical analysis
- experimental evaluation
- estimation algorithm
- learning algorithm
- matching algorithm
- selection algorithm
- segmentation method
- classification algorithm
- synthetic and real images
- convergence rate
- support vector machine svm
- similarity measure
- segmentation algorithm
- tree structure
- sparse matrix
- classification method
- objective function
- pairwise
- energy function
- probabilistic model
- machine learning
- calculation method
- input data
- expectation maximization
- em algorithm
- support vector machine
- optimal solution
- simulated annealing
- complexity analysis
- weight vector