On the reliable detection of concept drift from streaming unlabeled data.
Tegjyot Singh SethiMehmed M. KantardzicPublished in: Expert Syst. Appl. (2017)
Keyphrases
- concept drift
- reliable detection
- unlabeled data
- streaming data
- data streams
- semi supervised learning
- labeled data
- semi supervised
- co training
- class distribution
- active learning
- data distribution
- data points
- training data
- supervised learning
- learning algorithm
- labeled and unlabeled data
- non stationary
- text classification
- training set
- text categorization
- labeled examples
- change detection
- classification algorithm
- class labels
- sliding window
- prior knowledge
- positive examples
- data sets
- training examples
- feature extraction
- nearest neighbor
- training samples
- unsupervised learning