The benefits of word embeddings features for active learning in clinical information extraction.
Mahnoosh KholghiLance De VineLaurianne SitbonGuido ZucconAnthony N. NguyenPublished in: CoRR (2016)
Keyphrases
- information extraction
- active learning
- feature extraction
- question answering
- co occurrence
- image features
- natural language text
- machine learning
- free text
- feature set
- random sampling
- named entity recognition
- named entities
- linguistic knowledge
- data sets
- conditional random fields
- lexical features
- n gram
- text mining
- classification accuracy
- low level
- training set
- natural language
- training data
- decision trees
- learning algorithm