Interests should be able to be co-related with varying strengths, and thence there should be fascinating emergent properties.
By
Vynce
on February 08, 2007
For instance, an interest in video games is highly related to an interest in computer games (let's call it a .9 identity), and thus things tagged as related to computer games should also be considered (with a .9 strength) related to video games. On the other hand, an interest in travel might imply only a weak interest (.3) in foreign countries. Some things are stronger one way than another, perhaps, such as an interst in guitar might imply a strong (.8) interest in music but an interest in music might imply only a weak (.1) interest in guitar.
Agreed
24
1
96.0% 4.0%
Embed Claim
Make a related claim
Similar claims
- 30-0 Jyte should autosuggest creds based on the interests of both creditor and credited
- 18-0 There should be a single underlying object ("Subjects"?) used for implementing Creds, Interests, and Tags.
- 33-0 Jyte should enable filtering for claims tagged with your interests
- 9-7 Jyte should use the most popular capitalization for a particular tag
- 9-0 Cred dots should be filterable by tags
Discussion (5)
Very interesting. :)
How do you propose the relative correlation strengths are determined?
by claims, of course.
c.f.
http://jyte.com/cl/there-should-be-a-single-underlying-object-subjects-used-for-implementing-creds-interests-and-tags
and anent the current best word form question, of course i would tend to associate a thing with its plural (or other obvious stems) very highly.
also, this is a situation that could benefit from Oscar's Numerical Assessment Notation
Sure. I'll just have to learn R, I suppose, so the data analysis can really commence.
Claims inspired by this comment
Jyte should use R-project on Ruby on RailsR looks neat, but i'm not sure you need it for what i had in mind.
for making pretty pictures out of all sorts of random jyte data, though... that's cool.
Claims inspired by this comment
I want more pretty pictures of Jyte data