Semi-supervised training data selection improves seizure forecasting in canines with epilepsy.
Mona NasseriVáclav KremenPetr NejedlyInyong KimSu-Youne ChangHang Joon JoHari GuragainNathaniel NelsonEdward E. PattersonBeverly K. SturgesChelsea M. CroweTim DenisonBenjamin H. BrinkmannGregory A. WorrellPublished in: Biomed. Signal Process. Control. (2020)
Keyphrases
- semi supervised
- training data
- labeled data
- supervised learning
- unlabeled data
- eeg signals
- semi supervised learning
- training set
- multi view
- decision trees
- labeled examples
- data sets
- active learning
- human subjects
- short term
- unsupervised learning
- test data
- training process
- eeg data
- pairwise constraints
- epileptic seizures
- test set
- training examples
- ground truth
- support vector machine
- classification accuracy
- prior knowledge
- semi supervised classification
- learning algorithm
- machine learning
- class labels
- noisy data
- pairwise
- support vector
- labeled and unlabeled data
- video recordings