Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Build an algorithm to predict friendships, then actually use it to meet people

joingrouper.com

Post reply on HN
Print view View original

31–40 of 40 posts

Re: Build an algorithm to predict friendships, then actually use it to meet people

#31

Post by ergest »

mkwng wrote: post #22
streptomycin wrote: post #21

Earlier quoted context omitted.

Is there more description of the data anywhere? Like what does having an "f_number_of_pets" of 7.5 mean?

I just noticed in the FAQ it states, "...several fields have been renamed of course." If I'm understanding this correctly, any real-world conclusions you draw will be completely meaningless, as we're essentially working from a mislabeled dataset.

Not necessarily. They might as well be named attribute_1, attribute_2....attribute_n. ML algorithms don't care about the meaning of the features.

Re: Build an algorithm to predict friendships, then actually use it to meet people

#32

Post by chegra »

avalaunch wrote: post #9

This is a cool challenge but the prize is definitely lacking. I think anyone capable of writing an algorithm of the caliber you're looking for isn't likely to participate. I could be wrong but I think you're going to have to pony up some serious cash to get developers taking this seriously. Or you could go the more standard route and just hire someone to do the job.

I would do it cause I'm curious.

Edit: Now looking at the dataset, I wouldn't be able to use the model I developed personally.

Re: Build an algorithm to predict friendships, then actually use it to meet people

#34

Post by JFoss117 »

ergest wrote: post #31
mkwng wrote: post #22

Earlier quoted context omitted.

I just noticed in the FAQ it states, "...several fields have been renamed of course." If I'm understanding this correctly, any real-world conclusions you draw will be completely meaningless, as we're essentially working from a mislabeled dataset.

Not necessarily. They might as well be named attribute_1, attribute_2....attribute_n. ML algorithms don't care about the meaning of the features.

That's true, but to have the best chance of designing a good method/analysis, I need to know what the variables in my analysis mean. Otherwise, it is tougher to make decisions about what variables it makes sense to include in a model, what sorts of transformations make sense, what sort of approaches might work best, etc.

Re: Build an algorithm to predict friendships, then actually use it to meet people

#35

Post by JFoss117 »

nottombrown wrote: post #30
maxk42 wrote: post #26

I'll be the first to say it: Your data is either incorrect, arbitrary, or we're missing some information here. Why does everyone have "7.5" - 8 siblings and 7.5 - 8 "weekly workouts" and 7.5 - 8 platinum albums?

The headers with asterisks are intentionally mislabeled. Updated this to be more clear in the README.

You write in the README that the mislabeled columns are "from our internal ratings". Can you give any more definite sense of what this means? What kind of things are these ratings based off of? What are they designed to reflect? How are they computed (roughly)?

Re: Build an algorithm to predict friendships, then actually use it to meet people

#36

Post by idm »

JFoss117 wrote: post #34
ergest wrote: post #31

Earlier quoted context omitted.

Not necessarily. They might as well be named attribute_1, attribute_2....attribute_n. ML algorithms don't care about the meaning of the features.

That's true, but to have the best chance of designing a good method/analysis, I need to know what the variables in my analysis mean. Otherwise, it is tougher to make decisions about what variables it makes sense to include in a model, what sorts of transformations make sense, what sort of approaches might work best, etc.

I would echo this sentiment. Not only are the columns intentionally mis-labeled but they also appear to be computed, meaning some of the variance inherent to the original sample will have been lost.

Re: Build an algorithm to predict friendships, then actually use it to meet people

#38

Post by chegra »

Mutual Information for the fields:

I(f_facebook_friends_count,members_became_friends) = 0.117320113379

I(m_facebook_friends_count,members_became_friends) = 0.113972809724

I(m_facebook_photos_count,members_became_friends) = 0.0449092782303

I(f_facebook_photos_count,members_became_friends) = 0.0426531483254

I(m_shoe_size,members_became_friends) = 0.00276175766018

I(m_height,members_became_friends) = 0.00255043390135

I(f_shoe_size,members_became_friends) = 0.00233148724025

I(m_age,members_became_friends) = 0.00198005768283

I(f_height,members_became_friends) = 0.0013606978915

I(m_weekly_workouts,members_became_friends) = 0.00123271513215

I(f_age,members_became_friends) = 0.00122660347743

I(m_platinum_albums,members_became_friends) = 0.00111710129455

I(f_number_of_pets,members_became_friends) = 0.00108593667378

I(f_pokemon_collected,members_became_friends) = 0.000880040104571

I(m_number_of_siblings,members_became_friends) = 0.000830295252089

I(f_platinum_albums,members_became_friends) = 0.000820683185117

I(m_number_of_pets,members_became_friends) = 0.000768855827053

I(m_pokemon_collected,members_became_friends) = 0.000720822383999

I(f_weekly_workouts,members_became_friends) = 0.000620666529567

I(f_number_of_siblings,members_became_friends) = 0.00019278884716

I(f_gender,members_became_friends) = 0.000124279429698

I(m_gender,members_became_friends) = 0.000124279429698

Post reply on HN