Active learning for sound event classification by clustering unlabeled data.
Shuyang ZhaoToni HeittolaTuomas VirtanenPublished in: ICASSP (2017)
Keyphrases
- unlabeled data
- active learning
- labeled data
- semi supervised
- supervised learning
- semi supervised learning
- labeled and unlabeled data
- co training
- class labels
- unsupervised learning
- semi supervised classification
- data points
- text classification
- improve the classification accuracy
- training set
- selective sampling
- supervised learning algorithms
- machine learning
- labeled instances
- semi supervised learning algorithms
- unlabeled examples
- learning algorithm
- labeled examples
- random selection
- decision boundary
- clustering algorithm
- labeling effort
- label information
- unlabeled instances
- text categorization
- training data
- classification accuracy
- training examples
- pairwise constraints
- transfer learning
- supervised classification
- cost sensitive
- learning tasks
- supervised and semi supervised
- feature space
- class distribution
- random sampling
- k means
- sample selection
- multiple instance learning
- positive examples
- machine learning algorithms
- transductive learning
- feature selection
- feature extraction
- query by committee
- decision trees
- support vector
- training samples
- data mining