Tip:
Highlight text to annotate it
X
So basically, in this research we look at online dating websites.
We want to provide
some type of recommendation system for the dating website
so that when you visit the website you can find
recommended partners, we want to make sure
these are the users you are like, and then we also want to increase the chances
that these recommended users will like you back.
A lot of these dating websites
pick recommendations based on profile information.
For example, what you have put in your profile
including your race, your age,
your education level, whether you drink, whether you smoke,
and what kind of special characteristics you are looking for
in a partner.
And then the problem is, there are just so many questions out there,
just so much information, and people can sometimes get overloaded.
And sometimes a decision to pick a partner is actually a compromise.
When it comes to an
online dating website it's really about the data, about each user's profile,
as well as their historic activities, so
we believe that by analyzing these activities,
these previous activities, we can do a better job when it comes to recommending
a potential dating partner.
The underlying concept is actually
very similar to Netflix, right, but if you think about Netflix, you have the user on one side
and you have the movie on the other side. Users pick movies but
movies don't pick users. So, to to make this work in the online dating scenario we have
to make a lot of tweaks.
I'm still old-fashioned, I still
believe that to really find a soul mate, online is not enough. You still have to go to meet
the person
and talk to the person and decide. So, when it comes to
finding the perfect match, I will say
follow your heart, not only the algorithm online.