ContextGen: Targeted Data Generation for Low Resource Domain Specific Text Classification.
Lukas FrommeJasmina BogojeskaJonas KuhnPublished in: AISTATS (2022)
Keyphrases
- data generation
- text classification
- domain specific
- co training
- data streams
- general purpose
- machine learning
- high throughput
- text categorization
- feature selection
- naive bayes
- text mining
- labeled data
- active learning
- streaming data
- semantic features
- co occurrence
- feature extraction
- k nearest neighbor
- knn
- classification accuracy
- information retrieval