I am working on a dataset for an NLP task connected with sentiment analysis. While thinking on word roles in composing of the overall sentiment and marking particular as being positive/negative, we came across to three different situations:

  • word is considered generally positive: food, sleep, health, chocolate, help, politeness, etc;

  • word has an appraisal role: respectful, beneficial, responsible, objective (adj.), strong, useful;

  • word is emotionally charged: unforgettable, to admire, happy, magnificent, impressive.

Are there papers and/or research on the subject? I am looking for a formalized differentiation between those classes of positive/negative words.

  • I don't know an answer to your question as I'm unfamiliar with sentiment analysis, but would you consider these discrete categories, or are they on a continuous scale? – WavesWashSands Nov 19 '17 at 15:47
  • @WavesWashSands I have observed those categories having quite a solid boundary between each other. But within each category you can definitely put them onto a continuous scale. – Denis Kulagin Nov 20 '17 at 9:07

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.