Two Will Do: CNN With Asymmetric Loss, Self-Learning Label Correction, and Hand-Crafted Features for Imbalanced Multi-Label ECG Data Classification.
Cristina Gallego VázquezAlexander BreussOriella GnarraJulian PortmannGiulia Da PoianPublished in: CinC (2021)
Keyphrases
- multi label
- class labels
- multi label classification
- data sets
- classification accuracy
- multi label learning
- multiple labels
- image classification
- data analysis
- feature extraction
- text categorization
- hand crafted
- feature set
- text classification
- data points
- prior knowledge
- feature selection
- feature vectors
- image annotation
- feature space
- training set
- machine learning
- training data
- linguistic features
- support vector
- object recognition
- classification algorithm
- training examples
- knowledge discovery
- training samples
- labeled data
- model selection
- graph cuts
- co occurrence
- supervised learning
- knn