Non-parametric bounds on the nearest neighbor classification accuracy based on the Henze-Penrose metric.
Sally GhanemErik SkauHamid KrimHamilton Scott ClouseWesam SaklaPublished in: ICIP (2016)
Keyphrases
- nearest neighbor
- classification accuracy
- training set
- distance function
- triangle inequality
- finite sample
- k nearest neighbor
- knn
- instance selection
- nearest neighbor classification
- high dimensional
- naive bayes
- nearest neighbor search
- upper bound
- high dimensional data
- feature selection
- lower bound
- decision trees
- metric space
- training data
- support vector
- nearest neighbor algorithm
- metric learning
- accurate classification
- data sets
- reverse nearest neighbor
- probability density function
- supervised learning
- similarity metric
- feature space
- expected error
- nearest neighbor queries
- generalization bounds
- improvement in classification accuracy
- kernel density estimation
- upper and lower bounds
- classification rules
- distance metric
- euclidean distance
- cross validation
- distance measure
- lower and upper bounds
- vc dimension
- uniform convergence
- multi dimensional
- support vector machine