search

twitter / the-algorithm

Source code for Twitter's Recommendation Algorithm
https://blog.twitter.com/engineering/en_us/topics/open-source/2023/twitter-recommendation-algorithm
GNU Affero General Public License v3.0
61.78k stars 12.17k forks source link

Overview of comments sentiments #1759

Open darkdevildeath opened 1 year ago

darkdevildeath commented 1 year ago

The field that gives us engagement data could also use natural language processing and sentiment analysis to generate statistical data on positive or negative comments.

Influential people who receive a very large volume of comments can have a hard time understanding the type of comments they are receiving from their users and adjusting their content accordingly to these comments precisely.

The system could use natural language processing and sentiment analysis to identify what kind of posts the user has made and what the user is advocating for, and generate statistical data that shows whether the comments are positive and in agreement or if they are negative and in disagreement.

It is important to highlight that if a user says that a politician did something unpleasant and the comments also disapprove of the politician, the system should understand this as a positive comment and in agreement. Because, although the comments are negative for the politician, these comments are positive for the user.

nobodyindustries commented 1 year ago

The system could use magic dust, but it does not.

sunnamed434 commented 1 year ago

Well, I heard that the magic dust is on backorder, so I guess we'll have to settle for natural language processing and sentiment analysis for now. But who knows, maybe we can sprinkle some fairy dust on the algorithms and see what happens

aloiscochard commented 1 year ago

Duplicate of: https://github.com/twitter/the-algorithm/issues/1760

sunnamed434 commented 1 year ago

Duplicate of: #1760

that's means we need it!