• I wonder if keyphrase extraction by RAKE and short sentence extraction are related.

  • Key phrase extraction can be divided into two phases

    • Creating keyphrase candidates
    • Scoring key phrase candidates
  • I think concentration (of one’s attention) could be used for scoring.

  • When RAKE is compared to CRF, the “two or more occurrences” feature is different from a simple CRF [Global Features

  • Order by set size of candidate key phrases

    • explicit human-powered set
    • noun phrase
    • stop-word increment
    • RAKE-like coupling
    • substring
  • Order by length of key phrase candidates

    • character
    • word
    • stop-word increment
    • RAKE-like coupling
    • punctuation marks
    • short piece of writing (e.g. passage, article, composition)

This page is auto-translated from /nishio/キーフレーズ抽出20200823 using DeepL. If you looks something interesting but the auto-translated English is not good enough to understand it, feel free to let me know at @nishio_en. I’m very happy to spread my thought to non-Japanese readers.