Modifying kernels using label information improves SVM classification performance.
Martin Renqiang MinAnthony J. BonnerZhaolei ZhangPublished in: ICMLA (2007)
Keyphrases
- svm classification
- label information
- support vector machine
- semi supervised
- dimensionality reduction
- kernel function
- support vector machine svm
- training data
- support vector
- distance metric
- svm classifier
- labeled data
- classification accuracy
- class labels
- unlabeled data
- image classification
- multiple types
- classification method
- machine learning
- transfer learning
- semi supervised learning
- real world
- unsupervised learning
- geometric structure
- support vectors
- image processing
- multi class
- high dimensional
- pairwise
- feature extraction
- face recognition
- kernel methods
- hyperplane
- k nearest neighbor