Recurring Drift Detection and Model Selection-Based Ensemble Classification for Data Streams with Unlabeled Data.
Pei-Pei LiMan WuJunhong HeXuegang HuPublished in: New Gener. Comput. (2021)
Keyphrases
- concept drift
- model selection
- drift detection
- unlabeled data
- data streams
- ensemble classifier
- labeled data
- semi supervised learning
- semi supervised
- cross validation
- active learning
- hyperparameters
- machine learning
- text categorization
- data points
- training data
- ensemble learning
- supervised learning
- learning algorithm
- data distribution
- text classification
- unsupervised learning
- sample size
- regression model
- change detection
- training set
- feature selection
- prior knowledge
- data sets
- class labels
- generalization error
- non stationary
- classification algorithm
- transfer learning
- training examples
- learning problems
- learning models
- feature set
- high dimensional
- target domain
- pairwise
- pattern recognition