Reciprocal Collaborative Filtering

[funbobby]

Y'all seem like a pretty bright bunch... anybody here understand how the concept of collaborative filtering works and want to explain the nuances to me? If you don't know it by name, it's what Netflix and Amazon.com use to "recommend" items like movies and books to users. Anyways, I just found a nifty little site that uses the concept and applies it to Online Dating - you rate members and the system uses collaborative filtering to determine who you will like the most (seemed fairly accurate to me). Then, they take the idea one step further and tell you who is interested in YOU by using "reciprocal collaborative filtering". Take a look and tell me what you think- I think it's pretty revolutionary...

www.reciprodate.com

Btw, everything there was still free when I checked it out, so I suggest you create a user and rate some people and you'll see what I mean.

 

[phatslappy]

i dunno about "reciprocal" collaborative filtering, but the concept of regular CF has been around for quite some time. yes, you are correct that netflix and amazon use it. it is based on the idea of "similarity in tastes", meaning that if you and person X both like the same types of movies, and person X likes movie A, then chances are pretty good that you will like movie A. chances become even better that you will like movie A if person Y likes movie A, and you and person Y also have a high correlation in the types of movies you like. and so on...

i guess applying this concept to human dating makes sense. using the idea "reciprocally" is also quite interesting, since human beings are the only rateable items capable of rating back. if used correctly, this can potentially be an extremely powerful tool. i'm gonna go check out "Reciprodate" right now and see what i thing... hey why not, i'm single