vishalbhalla / Twitter-User-Personality-Prediction

Predict Personality of a person using Sentiment Analysis & Unigram Words as features on user's Twitter data.
23 stars 10 forks source link

Possible ways to improve Twitter #1

Open BradKML opened 3 years ago

BradKML commented 3 years ago

After reading https://www.demenzemedicinagenerale.net/pdf/2011%20-%20Predicting%20Personality%20from%20Twitter.pdf and others, there is a realization that this can be improved by reading response of Retweet/Reply.

  1. Defensive vs neutral tone to information (neurotic language)
  2. WIllingness to accept vs listen to information (agreeable language)
  3. Counter with new information vs emotional rhetoric (information keyword difference)
  4. Initiative of altering others vs one-on-one communication (@ others)
BradKML commented 3 years ago

BTW @vishalbhalla have you ever seen the Markov Chain-based Community Topic and Community Clustering?

Also Regarding Topic Sub-Networks https://sci-hub.se/10.3389/fdata.2019.00009 Issues: Hashtags are not a catch-all

BradKML commented 3 years ago

Another question: can N-grams be enough for Personality Detection? https://github.com/priyansh19/Classification-of-Personality-based-on-Users-Twitter-Data