Selecting Training Data for Unsupervised Domain Adaptation in Word Sense Disambiguation.
Kanako KomiyaMinoru SasakiHiroyuki ShinnouYoshiyuki KotaniManabu OkumuraPublished in: PRICAI (2016)
Keyphrases
- word sense disambiguation
- training data
- wordnet
- natural language processing
- machine translation
- data sets
- word sense
- wide coverage
- training set
- information extraction
- highly ambiguous
- lexical knowledge
- decision trees
- prior knowledge
- semantic relatedness
- linguistic knowledge
- unsupervised word sense disambiguation
- sense disambiguation
- classification accuracy
- feature generation
- learning algorithm
- machine learning methods
- semantic similarity
- part of speech
- semi supervised learning
- supervised learning
- text documents
- contextual information
- lexical information
- domain knowledge
- support vector
- semantic relations
- unlabeled data
- keywords
- knowledge base