Inside OKCupid: The math of online dating - Christian Rudder

by TED-Ed • 340,385 views

View full lesson: http://ed.ted.com/lessons/inside-okcupid-the-math-of-online-dating-christian-rudder When two people join a dating website, they are matched according to shared interests and how...

I think the assignment of the numbers up to 250 is the problem.  Think about it, all you would need is one or two "mandatory scores" and it would off set the rest of the things you DON'T have in common.  What if you had 5 questions that are all 'Very important' in common but you have two "Mandatory" questions that aren't in common out of 10 questions?  Assuming you got the same 'irrelevant' and 'somewhat'  shouldn't you be a decent match?  I mean not caring about a specific thing should be just as important as caring about it because it's still a personal value.  "Is Jay Z and Beyonce the best couple?"  A: Irrelevant, because I don't care.  B. Irrelevant because I don't know them.  People are asked BS questions (again a personal value) and if I remember correctly OkCupid allows it's users to create questions...that's a mistake because the number of BS questions and repetitive questions skews everything.  Two "Mandatory" questions that are similar would be answered in a similar way and throw the Algorithm results out of whack.  Humans are also prone to err, so we may be self-evaluating our values when we select 'how important' something is to us...we might even lie about it because it's easy to think better of yourself then be strictly honest.  Plus some people might change their answers thinking about the type of mate they want instead of just clearly explaining who he or she is.  Basically, too much human error and a lot of bad data leads to the unhappiness in the comments.
Nice to see something whose usefulness i wondered in high school... geometric mean in action ! Gives a whole new meaning to the burning question "how deep is your love?" ;)
I dated some guys off OkCupid who were like a 95-99% match with me and we all turned out to be ultimately incompatible, at least in the long term. Like my ex. Good enough to date casually but not relationship material. And I mean we both answered a lot of questions and I spent time going through his answers and explanations for them. Turns out he was just totally apathetic to people in general and devoid of true feelings. So eventually (after going on and off the site for some time. Some years) I was like screw this, and I deleted my account.
This video sums up what programming is about
Aren't people often bad judges of their inner desires and of what would actually work – i.e. what they think they would like in another person might not actually be a good predictor of compatibility? Because of this, I think I'd prefer an algorithm that also utilizes research on psychology of relationships and what kind of factors might predict happy long-term compatibility, statistically speaking etc.
you failed to mention that psychos and trannies sign up for this website and can hack your question system by lying. It happens a lot. People that sign up for these sites have a lot of issues, hence why they are on there.
"Pyschos and trannies"... Did you really just group those two in the similar categories?
Wait, what? Sometimes people dont even know what they want... Im very skeptical about this method
Wouldn't it be better to have an 75 percent match? I mean... If two people have too much in common it seems like there wouldn't be enough room to be an individual.
Only one problem, if someone lies about themselves. The computer can't pick out people who lie and no one wants to be with someone like that. But I don't really expect this problem to be solved. But I guess it could based on answers also calculate the likelihood of someone telling lies like if they rated themselves perfectly. Idk
It would be impossible to calculate the likelihood of someone lying without speculating. So really it's all still a gamble.
Carsten Kolassa Shared on Google+ · 7 months ago
The Math behind online Dating Christian Rudder one of the co-founders of OKCupid details how they calculate their match percentages. Nicely animated by TED-ED. Inside OKCupid: The math of online dating - Christian Rudder
OKCupid? Are these the people that started the avalanche that caused Brendan Eich to be hounded? I don't like.. http://www.forbes.com/sites/tonybradley/2014/04/05/backlash-against-brendan-eich-crossed-a-line/
Babak Keyvani Shared on Google+ · 6 months ago
An example of AI used in real-life scenario, and explained in plain English :)
Really, Online Dating is Actually Awesome. ( If you want to Online Dating with me then follow comment link.)
I have meet my current boyfriend on okcupid, we were a 71% match. I'm glad I joined.
I met my bf on ok cupid and we are only 66% match. I mostly use ok cupid for friends since I am new to the area and I like to meet different people.
still forever alone XD
There was some ugly women on ok cupid.  1 out of 100 were just ok.  plenty of fish has better looking women but online dating is messed up anyway.   Online dating is for low self esteem women.  Most women that are on there are looking just for messages to make them self feel better.  That is really what's going on
only gone in 1 minute of this video, and it's already so boring...
Were you expecting explosions or car chases or something?
+John Madden John Madden John Madden yes or something like how epipheo is doing (which I doubt you ever watched). Because they help promote people's startups through videos that is very engaging, fun and informative. While here, even though he's got a really nice animation, the speech and the voice just cannot gel with the motion graphics. And after a while it gets boring that it makes me want to stop watching and not care at all.
I participate in full-contact sports five nights a week.  No offense, but OKCupid keeps trying to match me to 300 lb. girls who pick four of their six "favorite things" as food items.  Think you need to check your algorithms a bit? 
Well nonetheless they are your "match" compared to your questions. You can always pay to have extra options to search though people too? i know one lets you only see "body type x"
I want to reply to this because I'm not sure that you took all the facts into account here.  You are an athlete correct?  So any question that asks something similar to "What is your diet like?", you will likely answer "I eat everything!" (Not actual questions and answer, simple example).  So what do you think the women you are being matched with might be answering as well?  Most of OkCupid's questions are personality based, math, religion, politics, sexuality, morals/ethics.  They algorithm is pretty good, you might just have to look a little deeper into the answers that you give compared to the results that you are getting.  Simply put, just because you are an athlete, doesn't mean that women with uncommon physique won't match with your personality.
Sadly, I get matched with guys that are lazy and just want me to massage their egos... Or complete weirdos... Who send me sexually explicit messages if I don't get back to them within 10 seconds. Okcupid is not for me. "The odds are good... But the gooda are odd!"
It is important to add another definition of social activity, occupation, nature of cognitive systems, date, place of birth and age - then it would be much more interesting.
Go out and meet people in REAL LIFE...people that use online dating websites are pathetic and losers!! So glad i met my girl in real life!
Do you want a medal or something? Some people have jobs or live in locations where it can be weird or awkward to just randomly meet someone new. By using a dating website, you ARE meeting someone new in real life. You have to plan to meet somewhere and talk face to face. You're only increasing your odds that you'll meet a good match, since in "real life" you're meeting completely random people, while on a website, you're meeting people with characteristics you'd like and who like your characteristics.
If u click on MATCHES section, it brings up mostly fat/ugly/not so good looking peoples. But when you click on LOCALS section, it brings up mostly cute/ beautiful/good looking peoples. OkCupid software seems to categorize the sections by peoples "looks" mainly.
+Andrew Friend Do a match percent search to at least sort them into an more or less stable order then start hiding the unattractive. I'm at 15k hidden users since April on my last account and have hit 30k on two prior accounts a few years apart. That's already about twice the people in my actual town given the recently inflated numbers. Good luck if you pass ~25, much less 30, though. You're even more fucked then.
okcupid taught me the value of hookers,and geometric mean.
okcupid you need to be pulled off the net.. you have fake profiles get fake people to reply to messages... all the women never respond to emails unless you are a 10... most if not all of them are rude.. most if not all are fat or ugly as fuck... you ban people aka guys only for reporting women that insult them even if the men are the victims.. you ban people if 10 women click block or 1 complains.. also you let it go for men if they are attractive.. i have proof 10 guy vs a 4 10 guy won out cuz you get money off clicks.. the site is a shit hole run by a fucking uk trash bastard that only cares about money.. they never personally answer emails only robots do.... the site along with match and plenty of fish are the same way.. wortthless
I don't understand the part where rooting the amount of questions, if there were 3 questions and there percentages are 91 and 98, it would 20%. 3√91+98 ≈  20%
You are abusing percentages.  3√91+98 ≈  20 ≈  2000%
I met my boyfriend on okcupid ( We are only a 48% match on there by the way, yet have a lot in common) but I must say a lot of guys on there are not looking for a serious relationship just a booty call and I don't believe the match percentage works because it's based on how many questions you answer, there are ALOT of questions on okcupid, more than 100! Most people, let alone guys, don't wait around to answer all the questions, therefore you cannot really say whether you are a good match or not. Not to mention people could totally lie with their response to the questions and you could be matched with someone you aren't really a match with. Sorry to put okcupid on blast but its true. Some questions are just stupid too. "Do you like to chew gum?" Seriously okcupid? What kind of question is that? LOL!
+tw19771 I think the algorithms work very well, that is why they are there. It isnt like some high school graduate is making them, they are mathematics lol It is true OKC is more of a booty call site, but their are is a large community using it for its intended us, to find new people they would get along with. You dont have to answer any of the questions if you dont want to. It is just another way to help match personalities   behind closed doors and in public.  The best thing about OKC is you dont get that bullshit in which you have to pay just to message someone or be able to read a profile. This is what OKC is doing right. and regarding the people making accounts as a joke, as long as they are not string someone on and lying to them making them waste their time, i dont have a problem with it; a lot of weirdo guys use it and even older men seeking/stalking younger women. 
I think the problem lies not with the program, but with the simple fact that people are douchebags.
All the math and scientific algorithms in the world don't mean shit if that other person thinks you're ugly. OKCupid's own stats show that people's match rates and who they respond to don't jibe and make no sense.
Obviously OkCupid works most accurately when users answer the questions thoroughly and truthfully. No, it does not weed out people who lie, which there is no way of preventing that.
1:47 He doesn't even mention the shit questions.
To put it bluntly and simply, it just doesn't that work that way. Like two people who like scary movies vs. one who does and one doesn't. You really ought to not out it into math for goodness sake.
    Tia, the point is not to say that people who have uncommon interests aren't attracted to each other.  The point to their algorithm is that people with more common interests are more likely to get along in an everyday interaction.  The movie example was simply that, it was an example.  I'm not sure if you have an OkCupid profile but they have thousands of questions.       What is worse, burning your countries flag or burning books?  Do you you put more weight into science or faith?  Is art important to you?  For me, I would answer burning your countries flag is worse, but it isn't that important to me.  I put more weight into science, and that is very important to me.  And art is important to me, because it is important to express ones self, but I don't practice any art forms myself, so on a scale this is not very important.  So you may look at those answers and say, wow, this guy would not be a good match for me, or vice versa.       OkCupid does that in their algorithm, but with hundreds and thousands of questions (depending on how many questions you answer).  Yes, just because you might like scary movies and  your partner doesn't might not be a deal breaker, however, if you are someone who enjoys watching scary movies every weekend, and your partner hates them, then that could potentially be an issue in a relationship.  Which is why they put the third portion into their algorithms, how important is this to you.  Hope this helps you understand the math behind the science of algorithmic love matching.
+Yuki Tokuda a I completely understand that. Honestly (this comment was so long ago I forgot what my cause for commenting was) my argument was directed more at the commenters of this video. Some of them really make me question people's intelligence. Of course you can successfully match people of similar interests. There's no question in that. Statistics can be used positively. But, the points in this video are too strong on one idea and to vague in another. I really don't care anymore. I had forgotten I even commented here.
lol. Christian Rudder. Muslim Anchor. Jewish Keel. Buddhist Bow. Hindu Knots? ehh, I ran out.
No, you got it wrong- Heisenberg and his wife are not that compatible.
I get zero girls to look at my profile, so I get very upset and suicidal when trying to find a girlfriend online. I tried take the best picture of myself as I can, it's just a very sad thing. I have gotten some fat girls to reply, but that's just about it. I hate online dating websites. I guess the only way to find a girlfriend is to have friends and go to dance clubs or something. I just have no hope, I just gotto commit suicide
Dude if you have the balls to kill yourself, you have the balls to fucking street approach.
LiveLeak Shared on Google+ · 6 months ago
Inside OkCupid: The math of online dating Check out the awesome animation +TED-Ed put together for cofounder Christian Rudder's lesson on how our matching algorithms work! When two people join a dating website, they are matched according to shared interests and how they answer a number of personal questions. But how do sites calculate the likelihood of a successful relationship? Christian Rudder, one of the founders of popular dating site OkCupid, details the algorithm behind 'hitting it off.' Source: http://ed.ted.com/lessons/inside-okcupid-the-math-of-online-dating-christian-rudder
The only things that matter to attract "slutwhores" is Looks & Money. My opinion.
why would you give away your business secrets, i can copycat you now
Statistical analysis isnt a secret. It is quite simple what they do; it isnt like they gave you the whole equation and outlier percentage. 
Geometric Mean. Something new I learnt today, from OKCupid dating algorithm.
somehow its so hard to date with someone on okcupid but great idea, thanks.
C4742 Shared on Google+ · 6 months ago
Genial simple Ähnlichkeitsberechnung. Vom schrägen Video nicht irritieren lassen.
all of that and okcupid never worked for me.. even if I match with another person 101% they simply won't reply to my messages.. all dating websites are useless no matter how good their algorithm is.. What I've realized is that it's really a matter of luck, it's possible to find a soul mate from another website that doesn't even have anything to do with dating
It is true though that looks are taken into a majority of someones decision to even click on your profile. I think %s are given too much power, it should simply be a helping hand to weed out people you wouldnt like. It is hard to get a reply, but you have to remember some profiles can be years old/ inactive. And sometimes they have just too many messages to look through. 
DatingFactoryFrance Shared on Google+ · 9 months ago
Date de OKCupid Matching La Rencontre en ligne est maintenant une industrie $ 2,000,000,000 . Merci à la croissance de ces sites comme Match.com, eHarmony, et OKCupid, l'industrie a augmenté de 3,5% par an depuis 2008 Les analystes s'attendent à cette accélération se poursuive au cours des cinq prochaines années - et pour une bonne raison:. C'est un très efficace moyen pour les couples de se rencontrer. Non seulement les sites de rencontres résultat dans les mariages les plus réussis , ils font un excellent travail correspondant à des couples potentiels en fonction de leurs diverses préférences et tendances. Et bien sûr, tout cela assorti est fait par des algorithmes.
only if the 80% of questions didn't require quite active and way open varied sex life.
Sure, that matches people.  But they need to take the results of the answers and compare them to the real world results of the dates they went on (a post date questionnaire will suffice) and see how successful the algorithm really is.
This totally goes hand in hand with another TED Talk about hacked online dating by a Jewish woman. I feel like what he was sharing here is that online dating uses algorithm to match people, taking OKCupid as an example. Not to explain every little detail about how they use the data. Also, I think everyone needs to keep in mind that people use dating sites because they are not meeting quality mate in real life (or maybe getting socially lazy to meet people in real life??). So, if we have to ask the computer/system to help with match making, this is how they can do it. It's a tool to help someone meet his/her mate who is also looking, and instead of talking to EVERY candidate, computer helps eliminate those who you tend to not "get along with". All in all, these videos are here so you can learn more about how stuff works, and you can make your own judgement if it works or not.
Yet for all that statistical calculation, if both people don't have at least an, "eh, I suppose," reaction to seeing the other's picture, it's all useless. Met some decent people, but none I was excited to see.
This is just an advertisement for OKCupid.
It might work better if anyone cared at all about the matching algorithm. Ultimately, it should have some predictive capability, but it doesn't. I have observed no correlation between match percentage and response rate. The only difference is that it is more of a disappointment to be ignored by a 96% than a 35%, though it happens no less often. Furthermore, this inorganic system cannot provide grounds or a subject for interaction; you can't open a message with a reference to how high your match percentage is, unless you couch it in a joke. As a result, those "superficial" items like common media favorites and shared interests become the crutch again, and since continued exchange relies on the success of those original interactions, they are significantly more important factors than the match percentages, which really will only predict the nature of an existing relationship, not the likelihood of one forming.
+Mikal Salaam Well, what's the problem? If it's a deal-breaker for superficial women out there, that's great! You now have an excellent filter to get rid of anyone not worth your time.
+Libertarian Prepper Yeah, I know, but the process is still frustrating and more than a little dispiriting.
My best friend met his wife on okcupid. Myself, I like to look at how a potential match answered a certain question prior to even contacting them. So I love okcupid as my favorite dating website. It's not the biggest, but it's by far the best. .
Dear voice-over author, The word, "data," is the plural of datum. "Data ARE used..." Not, "data is used..."
+Zev Eigen I agree with everything you wrote. I put info in quotes because I wasn't exactly sure if that was the right word to use... I erred on the side of regarding it as colloquial lingo. Second, your point about replicating ignorance is correct but I would argue that in the case of "data", as used in the video, you are too late. Data as a singular noun to refer to any bit of information, particularly the digital kind, has been acceptable usage for a very long time. I would argue that since the general population is not aware of the existence of the word datum, it is more efficient to use data as a singular noun that refers to variable amounts of information, regardless of the size or number of items therein. I guarantee that I am a fairly well-read and I still managed to evade datum till a few months ago. In other words, "data" may be used as a synonym/replacement for "information". You seem to be adamant in your view so I will leave it at this, but your "than me" example is a far cry from the issue of malapropism. Saying "better at grammar than me" is a logical nightmare. :)  http://en.wiktionary.org/wiki/data#Noun , Noun 2
I think I just get a little cranky after reading 2 many txts & twitters and such.
Then the only problem is not knowing what you really want or wanting something but not knowing what you need.  Also with questions like organization. One might think of themselves as average but another person my think they are messy or really neat so perceptive is also a flaw. I don't think you can math those out though so this seems reasonable enough.
People change everyday and to judge them on what they said at one point and time is wrong
+Barbadensis salicylate that's true, but within a single day, you can change responses to questions asked on the site, and you can constantly update the profile.  they took that into account
Using algorithms for matchmaking is like using horoscopes or numerology: it doesn't really work, because attraction is purely physiological. That's why some people who are a bad algorithm match have happy, lifelong marriages. They use the physiological attraction to justify getting along, despite their personality differences. Marriage is always work, because perfect psychological matches are extremely rare. if your algorithms match you to someone perfectly, that means absolutely nothing if there's no physiological basis in attraction. It all comes down to genetic compatibility, and scent, auditory, tactile and visual markers that our brain is hard-wired to interpret as evidence of genetic compatibility. Anything beyond that is just superficial similarities and differences. This is true even in homosexuals, albeit in a heterosexually simulated way (i.e. they would be a perfect breeding match if one were a woman and the other a man).   Dating algorithms are a scam. Case in point: I tried eHarmony, and they matched me with some chick from my childhood hometown, who used to bully me in grade school and junior high school. She grew up to be a type A personality, just like me. How in hell they matched us up, I don't know, but even if there wasn't the baggage issue, two type A's together spells disaster. Furthermore, I was physically repulsed by her when I knew her.  Anyone reading this, don't waste your money on sites that use algorithms to hook people up. They all say they're the most successful site with the most matches ending in marriage, but that's pure bullshit. You have a higher chance of finding marriage material through personal introduction or casual interactions in meat-space.
Real scientists citate peer reviewed articles. Don't pretend you're talking about science when you're talking about your own extrapolations and conclusions about some study or about your own mind. If you have conclusive evidence to prove your theories, present them. Otherwise, mark your thoughts with "I think that" or "I don't think that".
Women are all the same upside down.  It was the same for Trixie.  She liked it rough. 
good luck with your love life.
Total bullshit.  OKCupid's geeked out algorithms usually don't work.  It's a gimmick and they are no better than any other dating site.  This is BECAUSE people are making judgements on how they like a person's PICTURE, their RACE/COLOR, and socio-economics.  Furthermore, people can give you politically and ethically correct answers to questions, but in actuality they BEHAVE in ways that CONFLICT or are the OPPOSITE of their answers. Lastly, many people view matching percentages like test scores.  So they assume 90% is better than 65%.  This means people (primarily men) will game the system specifically to get a HIGHER matching percentage, in CONFLICT with their actual personality.  And many people (often women) will read too much into a guy having a high match percentage or arbitrarily dismiss guys with lower matching percentages.
Gaylene Brazill Shared on Google+ · 1 year ago
Is DateFindersk effective to met your girlfriend? We have read many good things about DateFindersk (google it).
The scale/increments for how important a question is are skewed in my opinion.   Not very important = 1 while irrelevant = 0 (both on a 250 point scale).  Seems like you're trying to boost % matches...
Some of my best friends are the people I have less in common with.
est naiviv Shared on Google+ · 1 year ago
Yuting Pang Shared on Google+ · 2 years ago
Christian Rudder explains how the dating algorithm works:
I love online dating but I have very little faith in algorithms. Some things are best left to the individuals involved.
You`re so right! Let´s let them do there work and enjoy getting matched with all those mr. rights out there :-)
Joe Frawley Shared on Google+ · 2 years ago
Neil Tamzali Shared on Google+ · 2 years ago
L'algorithme d'OK Cupid raconté par son fondateur. 100% pédagogie.
Sian Liu Shared on Google+ · 2 years ago
Oshyan Greene Shared on Google+ · 2 years ago
Ever wondered how OK Cupid's matching algorithm works? How about a real, actual math breakdown? Yes. (unfortunately nothing on "friend" and "enemy", but still interesting) Inside OKCupid: The math of online dating - Christian Rudder!
John Oliver Shared on Google+ · 2 years ago
Abraham Samma Shared on Google+ · 2 years ago
How does a computer match a pair for an online date? With a love algorithm of course!
Juggler TS Shared on Google+ · 2 years ago
The Math of Online Dating A TEDEd talk from one of the founders of OKCupid explaining how OKCupid's matching algorithm works. It's very very simple! And here's a similar video from the other day: The Odds of Finding Life and Love | It's Okay to be Smart | PBS Digital Studios
Matching Books to Buyers It's actually not that dissimilar to love in a lot of ways. I found this video about how OkCupid does their matching algorithm very thought provoking. Here's where it led me: 1) What if you could answer questions about your story premise and get a numerical result of possible buyers that would by highly satisfied with that premise? 2) Would you take the time to do the same thing for your book upon its release and trust the algorithm to handle discoverability? 3) As a reader, would you pay a small premium to be recommend books you have a high likelihood of enjoying?
Perhaps it could operate as Pandora does.  I'd be interested as a reader to find books that have certain qualities and be able to exclude others.  For instance, I'm interested in love stories (aka romance) but I keep finding stuff that pisses me off (rape, rapeyness, alpha jerks, etc).  If I could say I want something in Category: Romance but then go through a list of standard tropes to x-out ones that piss me off, that'd be cool. As a writer, I'd happily take the time to enter which pre-set things my book has.  However, I can do that easily enough as an indie now.  When I was with several small press publishers, none of them would have had the manpower to bother with that, and I'm not sure I would have trusted them to get it right anyway.  And as a writer I'd be peeved if my book ended up with a wrong label. So I think this is a rare concept that might work better for indies than for traditionally published books.
+Kimberly Chapman Very cogent points, though I feel that even Traditional Authors would likely take the time if it meant they could get quality matches. I agree though, I'm approaching this as an author first and a reader second. Though I see that it is valuable for both. More than x-ing out, I think you could also specify what you really enjoyed too, so not only would you find books free of things that piss you off, but you'd get higher rankings of books that have all the good stuff you like too. Pandora is an aggregate algorithm as well, learning more as you interact with it. Obviously people like that approach. It's a very good comparison. Thanks for your suggestions, they are very much appreciated.
Anna Jacobi Shared on Google+ · 1 year ago
Ilustrado muy claro: Christian Rudder de OKCupid nos explica la matemática del amor en línea.
Great educator... and entertaining also
Very interesting approach!
Col McColl Shared on Google+ · 2 years ago
The math of online dating
The part of his company that's worth money is his client base. Anybody can write a simple algorithm like this.
Why do you use the n-th root and not always the square root? You calculate the geometric mean of two numbers and not of n numbers (with n = number of commonly answered questions). With this algorithm, two people with match for A 70% and match for B 70% will get a common match of 95% if they have 15 commonly answered questions. Is that your intention? I think the common match should be 70%, too.
Too bad every1 isn't using matchmaking. Only the weird and somewhat pathetic people. No offence, it's just how society see them.
True - I am sure that one of the biggest sources of error in a system like this (ie. getting a high score with someone you don't end up liking when you actually date them) is that many people don't really know what they want. Their self-evaluation might be a ways off from what their reality is (like thinking you want to be the center of attention, when in reality you feel more comfortable when others take the lead). Still it is a useful tool even when it only works a little.
It is all about desire - for both men and women !!!
Very informative and math friendly video. Realitymalfunctions - I think this is just a trolling poster, or person trying to promote a site. Ignore them for best results.
I heard that OKCUpid at launch was not Gay or Atheist friendly. Is this true?
Not at all. Physical attraction is far more subjective, and there is no computer algorithm currently available that's going to predict what kind of photo of a man a woman will find physically attractive.
I had a date with a 94% or so match from OKCupid some time ago. We had a lot of common interests. We had read or wanted to read the same books, watched or wanted to watch the same movies, played or wanted to play the same games. We laughed about the same things. We had similar goals and aspirations in life. We had almost identical personalities. It was creepy. We never met again. True story.
The girl who did the animation Celeste Lai, is super hot.
they may include that to in the data collection phase, don't you think?
It would be interesting to see if this algorithm can measure the likelihood of two people getting together who gave completely opposite answers to the questions. It would cast a light on discovering whether opposites truly attract and with what frequency.
As long as they don't type using all caps and a language as poor as yours, I don't really mind feminist or bisexual girls.
Seriously! That's fucking crazy if it's true. That's halarious.
This is a schoolbook example of an expert system. I had a whole subject dedicated to such systems last semester. Awesome subject :)
True. Though I'm sure that could be factored in somehow as well. I know OKCupid let's ppl rate the attractiveness of members on a 5 point scale. So using historical data on how ur messaging habits are influenced by the attractiveness rank of the member ur messaging combines with the data of how other members have rated the attractiveness of members being matched up, I'm sure you could add in an educated guess for potential physical attraction as well.
Or you could just find someone in real life. LOL Pretty awesome algorithms though (y)
Show more Loading...
to add this to Watch Later

Add to