Paired compressed cover trees guarantee a near linear parametrized complexity for all k-nearest neighbors search in an arbitrary metric space.
Yury ElkinVitaliy KurlinPublished in: CoRR (2022)
Keyphrases
- metric space
- k nearest neighbor
- knn
- query point
- reverse k nearest neighbor
- distance function
- nearest neighbor
- similarity searching
- similarity search
- reverse nearest neighbor
- nearest neighbor search
- distance computation
- range queries
- similarity queries
- searching in metric spaces
- similarity search in metric spaces
- indexing techniques
- high dimensional
- support vector machine
- skyline queries
- locality sensitive hashing
- management system
- knn classifier
- r tree
- data structure
- access methods
- search methods
- query refinement
- linear systems
- data sets
- vector space
- text classification
- data points
- feature space
- decision trees
- feature selection
- databases