A kernelized non-parametric classifier based on feature ranking in anisotropic Gaussian kernel.
Razieh SheikhpourMehdi Agha SarramMohammad Ali Zare ChahookiRobab SheikhpourPublished in: Neurocomputing (2017)
Keyphrases
- gaussian kernel
- feature ranking
- diffusion equation
- support vector machine
- kernel function
- scale space
- feature selection
- anisotropic diffusion
- kernel methods
- partial differential equations
- diffusion process
- high dimensionality
- support vector
- noise removal
- reproducing kernel hilbert space
- feature subset
- machine learning
- text categorization
- multi class
- optimal parameters
- training data
- gaussian process
- missing values
- neural network
- edge detection
- generalization ability
- feature vectors
- knn
- image processing
- multiscale
- pairwise
- multi layer perceptron
- support vector machine svm
- k nearest neighbor
- learning tasks
- svm classifier