Something I want to figure out how to do is compact...

#Segmented#Hashtags#Maybe? 😜

8 Feb 2020 at 14:54 | Open on elekk.xyz

FiXato

@Gargron maybe check for an overlap in trending hashtags? Hashtags which not only look similar, but are also used together in multiple toots?

Alternative could be for instance admins to have an option to group tags together manually, and for users to perhaps suggest tags to be grouped?

8 Feb 2020 at 14:56 | Open on mastodon.social

Cats Who Code

@Gargron Is there a way to redirect them, or something along those lines?

8 Feb 2020 at 15:00 | Open on mastodon.social

ÁlexMárquezPérezMuñízDíazPúras

@Gargron ¿Would it change things too much if we make searches partial-matching instead of whole-word or prefix?

8 Feb 2020 at 15:03 | Open on mastodon.social

Mr. Funk E. Dude

@Gargron how about a suggested hashtag based on key words in the most popular tags in that moment?

8 Feb 2020 at 15:06 | Open on mastodon.social

Paul

@Gargron Hell yes, one start would be collapsing plurals. Maybe a synonym dictionary would get you about half of the way there.

8 Feb 2020 at 15:30 | Open on mastodon.social

Trevor Meier

@Gargron I wonder if there’s some cross-pollination between how #SSB allows unique canonical users to be tagged “same-as”? In their case it allows a unique ID for eg mobile and desktop accounts to be understood as being from the same person. Something similar could work for hashtags

8 Feb 2020 at 15:40 | Open on social.coop

Jordan who has a paper due

@Gargron You could use something like worldnet. https://wordnet.princeton.edu/
Those two tags you gave would be very close together in terms of lexical semantics (each word has a similar meaning). Like if you look up "domain term categories" for election you can see how poll is related. http://wordnetweb.princeton.edu/perl/webwn?o2=1&o0=1&o8=1&o1=1&o7=1&o5=1&o9=&o6=1&o3=1&o4=1&s=election&i=2&h=10000000#c
If you could order tags in a way that total lexical distance in wordnet = distance in lexical rank. And then show tags that are within a range of lexical rank.

@Gargron You could use something like worldnet. https://wordnet.princeton.edu/
Those two tags you gave would be very close together in terms of lexical semantics (each word has a similar meaning). Like if you look up "domain term categories" for election you can see how poll is related. http://wordnetweb.princeton.edu/perl/webwn?o2=1&o0=1&o8=1&o1=1&o7=1&o5=1&o9=&o6=1&o3=1&o4=1&s=election&i=2&h=10000000#c
If you could order tags in a way that total lexical distance in wordnet = distance in lexical...

Expand text...

8 Feb 2020 at 16:44 | Open on mastodon.social

Jordan who has a paper due

@Gargron You could use something like worldnet. https://wordnet.princeton.edu/
Those two tags you gave would be very close together in terms of lexical semantics (each word has a similar meaning). Like if you look up "domain term categories" for election you can see how poll is related. http://wordnetweb.princeton.edu/perl/webwn?o2=1&o0=1&o8=1&o1=1&o7=1&o5=1&o9=&o6=1&o3=1&o4=1&s=election&i=2&h=10000000#c
If you could rank tags in a way that total lexical distance would correspond to a higher distance in rank. And then show tags that are within a range of lexical rank.

@Gargron You could use something like worldnet. https://wordnet.princeton.edu/
Those two tags you gave would be very close together in terms of lexical semantics (each word has a similar meaning). Like if you look up "domain term categories" for election you can see how poll is related. http://wordnetweb.princeton.edu/perl/webwn?o2=1&o0=1&o8=1&o1=1&o7=1&o5=1&o9=&o6=1&o3=1&o4=1&s=election&i=2&h=10000000#c
If you could rank tags in a way that total lexical distance would correspond to a higher distance...

Expand text...

8 Feb 2020 at 16:50 | Open on mastodon.social