Random Forests with Random Projections of the Output Space for High Dimensional Multi-label Classification.
Arnaud JolyPierre GeurtsLouis WehenkelPublished in: ECML/PKDD (1) (2014)
Keyphrases
- random forests
- high dimensional
- input space
- multi label
- dimensionality reduction
- low dimensional
- dimension reduction
- high dimensional data
- high dimensionality
- class labels
- decision trees
- text categorization
- machine learning algorithms
- ensemble methods
- image classification
- feature space
- image annotation
- graph cuts
- kernel function
- hyperplane
- text classification
- data points
- learning tasks
- binary classifiers
- unsupervised learning
- binary classification
- feature selection
- k nearest neighbor
- principal component analysis
- pattern recognition
- feature extraction
- input data
- benchmark datasets
- training samples
- sparse representation
- prediction accuracy
- neural network
- image processing
- learning algorithm
- information retrieval