search

CodeforLancaster / sitegeist

4 stars 1 forks source link

Tweak algorithm to disregard any superfluous words #17

Closed neilmorton closed 3 years ago

neilmorton commented 5 years ago

When looking at http://www.codeforlancaster.org.uk at the CfL meet on 30th April 2019, words such as "who" were showing in the sentiment.

Once we have some data, perhaps any superfluous words can be added to a disregard list (which we assume must exist for other words already).

ryancallihan commented 5 years ago

I noticed the same thing. I will check that out along with making the "Entities" and "Phrases" a bit better.

ryancallihan commented 5 years ago

I made some tweaks. I am going to let it run a bit to see how much better it is working.

ryancallihan commented 5 years ago

Ok, I think I got something that works pretty well. I am going to submit a pull request for it. I added a "Words" and "Emojis" option as well.

neilmorton commented 3 years ago

Closing issue as group no longer active.