- plot out frequency of values each feature takes on.
- figure out the distribution for each random variable.
- Semantic disambiguation
- Semantic web
- Linked data?
- Ekaterina Shutova: https://scholar.google.ca/citations?hl=en&user=jqOFBGoAAAAJ&view_op=list_works&sortby=pubdate
-
Wordnet has api for classification of words by type, topic, abstractness. here: http://www.nltk.org/book/ch02.html
-
Ground level: bag of words comparison.
-
gensim api.
- knowledge graph
- word association (these words are found together)
-
what about: knowledge graph where each word has a subgraph of related words
- if a word seems out of place but the subgraph contains words that would make sense in the given context, it may be a metaphor
- use synonyms as the subgraph.
- delmonte paper on metaphor
- cowell and gerophski paper
-
Poets that favor certain senses over others
-
quantify degree of abstractness
-
Choice of words can give the poem a certain feel. Formal, casual, tense, playful.
-
How is poetic speech different from ordinary speech?