Using one-class classifiers and multiple kernel learning for defining imprecise geographic regions.
Eduardo CunhaBruno MartinsPublished in: Int. J. Geogr. Inf. Sci. (2014)
Keyphrases
- multiple kernel learning
- geographic regions
- data sets
- land cover
- linear combination
- location based services
- learning problems
- kernel learning
- kernel methods
- kernel function
- semi supervised learning
- binary classification
- class imbalance
- spatial data
- satellite images
- learning models
- multi task
- training data
- multiple features
- support vector
- image processing
- feature extraction
- user groups
- support vectors
- pattern recognition
- semi supervised
- multispectral