Employing the Principal Hessian Direction for Building Hinging Hyperplane Models.
Anca Maria IvanescuThivaharan AlbinDirk AbelThomas SeidlPublished in: ICDM Workshops (2012)
Keyphrases
- hyperplane
- feature space
- input space
- linear classifiers
- support vector
- kernel function
- training samples
- incremental learning algorithm
- support vector machine
- linearly separable
- principal components
- data points
- data sets
- locality sensitive
- support vectors
- kernel space
- svm classifier
- convex hull
- model selection
- prior knowledge
- decision trees