Fully Data-Driven Optimization of Gaussian Parameters for Kernel Classifier.
Kosuke FukumoriToshihisa TanakaPublished in: APSIPA (2018)
Keyphrases
- data driven
- maximum likelihood
- support vector
- kernel machines
- linear classifiers
- feature space
- training data
- fine tuning
- parameter values
- gaussian mixture
- normal distribution
- parameter estimation
- optimization problems
- kernel parameters
- rbf kernel
- parameter optimization
- optimal kernel
- multivariate gaussian distribution
- feature selection
- decision trees
- optimal parameter values
- kernel methods
- multiple kernel learning
- optimization algorithm
- probability density function
- optimization criteria
- training samples
- machine learning
- kernel classifiers
- learning algorithm
- gaussian kernel
- input data
- nonlinear functions
- class labels
- svm classifier
- training process
- gaussian distribution