Active learning for reducing bias and variance of a classifier using Jensen-Shannon divergence.
Minoo AminianPublished in: ICMLA (2005)
Keyphrases
- jensen shannon divergence
- selection criterion
- active learning
- bias variance decomposition
- feature selection
- information theoretic
- model selection
- variance reduction
- information theory
- low variance
- label noise
- learning algorithm
- bias variance
- mutual information
- training examples
- jensen shannon
- training set
- selective sampling
- intra class
- semi supervised
- sample selection bias
- machine learning
- support vector machine
- support vector
- decision trees
- training data
- random sampling
- classification accuracy
- variable selection
- naive bayes classifier
- classification models
- covariance matrix
- sample size
- machine learning algorithms
- active learning framework
- active learner
- data mining