Bucketing coding and information theory for the statistical high-dimensional nearest-neighbor problem.
Moshe DubinerPublished in: IEEE Trans. Inf. Theory (2010)
Keyphrases
- information theory
- nearest neighbor
- high dimensional
- information theoretic
- rate distortion theory
- high dimensional data
- nearest neighbor search
- statistical learning
- jensen shannon divergence
- statistical mechanics
- data points
- dimensionality reduction
- knn
- coding scheme
- relative entropy
- training set
- statistical physics
- low dimensional
- conditional entropy
- information geometry
- kullback leibler divergence
- feature space
- similarity search
- mdl principle
- mutual information
- shannon entropy
- data sets
- metric space
- variable selection
- decision trees