3 ms·

On page 3, One of the founders of a YC company called 1000Memories.com (it's FB for dead people, only more interesting) says he heard FB can already tell when

On page 3, One of the founders of a YC company called 1000Memories.com (it's FB for dead people, only more interesting) says he heard FB can already tell when you're about to break up with someone: certain communication patterns emerge. Yes, I know that the amount of data people share on Facebook is staggering, but, you've got to be kidding me!

hugh3 16y ago
"Can tell", surely, with some accuracy better than random but not nearly good enough for actual predictions. It would be interesting to know how the algorithm works, though... if only for my own personal usage. Is it a flurry of sudden communication? Or one partner starts ignoring the other? Do you need to know the words they're using, or can you just get it from the frequency of various types of communication? Facebook does scare me enough, however, that I refuse to be friends (on facebook or in real life) with anyone who works at facebook. When I meet someone socially who works at facebook (which has only happened a couple of times so far) my next reaction is to edge slowly away from the conversation. And you'd have to be crazy to date someone who works at facebook. And now I come to think of it, all the facebook employees I've met have been single...
ig1 16y ago
Nah, I'm pretty sure that's just a Chinese Whisper that originated from David McCandless work in data mining breakups on Facebook. It can tell you what the most popular day to breakup is, but it can't predict when a couple is going to breakup. However that said it probably is possible to predict breakups with communication data. You could probably even do it with an iphone app, breakups tend to follow certain events (like drop in respect, distancing, etc.) that can detected in vocal patterns.
- siddhant 16y ago
  There is a (YC?) company that analyzes your GMail and tells you which of your friends you haven't been in touch with, in recent times. I'm forgetting the name right now. But we all know that Facebook has an edge over GMail in terms of personal data, and they can certainly do it. Hmm. I can see an app waiting to be built there. :-)
  - shykes 16y ago
    You're thinking of Etacts.
    - siddhant 16y ago
      Ah yes. Etacts it is. Thanks!
rokhayakebe 16y ago
"certain communication patterns" == S/he emailed her friend "I think s/he is cheating, I dunno anymore".
woodall 16y ago
I just finished my internship for this semester. I had the opportunity to see how death cases are investigated. One case in particular involved cell phone records. After transcribing all of the calls into a ordered list a clear pattern emerged. When a crime is about to be committed, phones ring off the hook. This was confirmed by a state agent later. The idea that FaceBook can predict patterns in people's relationships with others is not all that surprising; especially since they have context. Language is a good predictor of many things. Check out 'Gender Genie', a free program to predict the sex of an author. http://www.bookblog.net/gender/genie.php http://www.bookblog.net/gender/genie.php
patio11 16y ago
I totally wouldn't put it past them, with the proviso that the proper claim is that they are able to predict when someone goes from "in a relationship" on FB to "single" or "in a relationship [with someone else" on FB, and this isn't perfect but it is much better than you'd expect from random chance. For example, I think that Facebook could identity a pool of 100,000 users tomorrow and predict a breakup within the next two weeks, and be right for over 20% of them. That is statistically improbable unless you expect relationships to average about 10 weeks. (Yeah yeah, handwavy math there. Wave wave.) How you'd do this? The same way you do any AI problem: cheat like a mofo until it looks like a problem we've already solved. For example, come up with a feature space of, say, 100 things which are easily algorithmically checkable: messaged boyfriend in past day, messaged boyfriend more than 10 times in last day, messaged boyfriend in last week, messaged boy other than boyfriend in last day, ..., changed sexual orientation listed in FB, changed residence listed in FB, ... went off to college, whatever. Assign each factor a random number from 1 to 100. Now, construct a Hadoop cluster, grab your favorite million Facebook accounts, and use any AI technique you like (simulated annealing, GA, whatever) to play with those scores until you find a decent fit where the dot product of that vector and a vector of binary tests against a user delivers high scores for people who break up and low scores for people who don't. Then, grab a hundred million Facebookers, take the dot product for each of them against your best performing vector, and print out the 100k with the highest scores. This will work. Indeed, it almost can't not work, if any of your constituents of the feature vector captured any useful information. (P.S. The experimental design here is ridiculously biased in favor of success. Real life sometimes is -- it is certainly true for FICO, for example, since success is "beats a human underwriter" and human underwriters suck, what with their sucky salaries and sucky accuracy and sucky inability to process a hundred thousand applications an hour.)

hugh3 16y ago

"Can tell", surely, with some accuracy better than random but not nearly good enough for actual predictions. It would be interesting to know how the algorithm works, though... if only for my own personal usage. Is it a flurry of sudden communication? Or one partner starts ignoring the other? Do you need to know the words they're using, or can you just get it from the frequency of various types of communication? Facebook does scare me enough, however, that I refuse to be friends (on facebook or in real life) with anyone who works at facebook. When I meet someone socially who works at facebook (which has only happened a couple of times so far) my next reaction is to edge slowly away from the conversation. And you'd have to be crazy to date someone who works at facebook. And now I come to think of it, all the facebook employees I've met have been single...

ig1 16y ago

Nah, I'm pretty sure that's just a Chinese Whisper that originated from David McCandless work in data mining breakups on Facebook. It can tell you what the most popular day to breakup is, but it can't predict when a couple is going to breakup. However that said it probably is possible to predict breakups with communication data. You could probably even do it with an iphone app, breakups tend to follow certain events (like drop in respect, distancing, etc.) that can detected in vocal patterns.

siddhant 16y ago
There is a (YC?) company that analyzes your GMail and tells you which of your friends you haven't been in touch with, in recent times. I'm forgetting the name right now. But we all know that Facebook has an edge over GMail in terms of personal data, and they can certainly do it. Hmm. I can see an app waiting to be built there. :-)
- shykes 16y ago
  You're thinking of Etacts.
  - siddhant 16y ago
    Ah yes. Etacts it is. Thanks!

rokhayakebe 16y ago

"certain communication patterns" == S/he emailed her friend "I think s/he is cheating, I dunno anymore".

woodall 16y ago

I just finished my internship for this semester. I had the opportunity to see how death cases are investigated. One case in particular involved cell phone records. After transcribing all of the calls into a ordered list a clear pattern emerged. When a crime is about to be committed, phones ring off the hook. This was confirmed by a state agent later. The idea that FaceBook can predict patterns in people's relationships with others is not all that surprising; especially since they have context. Language is a good predictor of many things. Check out 'Gender Genie', a free program to predict the sex of an author. http://www.bookblog.net/gender/genie.php http://www.bookblog.net/gender/genie.php

patio11 16y ago

I totally wouldn't put it past them, with the proviso that the proper claim is that they are able to predict when someone goes from "in a relationship" on FB to "single" or "in a relationship [with someone else" on FB, and this isn't perfect but it is much better than you'd expect from random chance. For example, I think that Facebook could identity a pool of 100,000 users tomorrow and predict a breakup within the next two weeks, and be right for over 20% of them. That is statistically improbable unless you expect relationships to average about 10 weeks. (Yeah yeah, handwavy math there. Wave wave.) How you'd do this? The same way you do any AI problem: cheat like a mofo until it looks like a problem we've already solved. For example, come up with a feature space of, say, 100 things which are easily algorithmically checkable: messaged boyfriend in past day, messaged boyfriend more than 10 times in last day, messaged boyfriend in last week, messaged boy other than boyfriend in last day, ..., changed sexual orientation listed in FB, changed residence listed in FB, ... went off to college, whatever. Assign each factor a random number from 1 to 100. Now, construct a Hadoop cluster, grab your favorite million Facebook accounts, and use any AI technique you like (simulated annealing, GA, whatever) to play with those scores until you find a decent fit where the dot product of that vector and a vector of binary tests against a user delivers high scores for people who break up and low scores for people who don't. Then, grab a hundred million Facebookers, take the dot product for each of them against your best performing vector, and print out the 100k with the highest scores. This will work. Indeed, it almost can't not work, if any of your constituents of the feature vector captured any useful information. (P.S. The experimental design here is ridiculously biased in favor of success. Real life sometimes is -- it is certainly true for FICO, for example, since success is "beats a human underwriter" and human underwriters suck, what with their sucky salaries and sucky accuracy and sucky inability to process a hundred thousand applications an hour.)