Influence of training sampling protocol and of feature space optimization methods on supervised classification results.
Sylvie DurrieuThierry TormosPascal KosuthCatherine GoldenPublished in: IGARSS (2007)
Keyphrases
- optimization methods
- supervised classification
- supervised learning
- feature space
- training set
- optimization method
- training samples
- simulated annealing
- unsupervised learning
- optimization problems
- supervised classifiers
- unsupervised clustering
- global convergence
- unconstrained optimization
- optimization approaches
- dimensionality reduction
- stochastic methods
- machine learning
- high dimensional
- training data
- quasi newton
- direct optimization
- bayesian network models
- evolutionary algorithm
- mean shift
- pairwise
- image segmentation
- unsupervised classification
- pattern recognition
- data sets
- information retrieval
- genetic algorithm
- feature selection
- data points
- hyperplane
- metaheuristic
- kernel function
- active learning
- cost function