Having a Blast: Meta-Learning and Heterogeneous Ensembles for Data Streams.
Jan N. van RijnGeoffrey HolmesBernhard PfahringerJoaquin VanschorenPublished in: ICDM (2015)
Keyphrases
- meta learning
- data streams
- base classifiers
- decision trees
- meta knowledge
- inductive learning
- learning tasks
- model selection
- machine learning algorithms
- ensemble methods
- streaming data
- concept drift
- data sets
- multi class
- data mining
- training data
- feature selection
- machine learning
- database
- training samples
- metamodel
- text categorization
- naive bayes
- prediction accuracy
- transfer learning
- learning algorithm
- real world