Monday, April 4, 2011

Use the Analytics to Find a Perfect Match

  It is known that internet dating has become lucrative multi-billion dollars industry.  Part of the service is helping date seekers to find “ideal” match. So how do they do that? Well, dating service companies like use sophisticated algorithms to profile its date-seeking base and then match them against similar profiles. But, it goes far more sophisticated than that. There are some new Facebook applications that create pseudo-DNA profile of its users in terms of preferences for music, hobbies, movies, as well as topical preferences using their Twitter threads and match them against each other to find a match.

Critics may say that very important dimension by using analytics and data mining to choose the match – is omitted. Dimension of physical appearance! Well , wrong! Some of these applications are able to do visual mining to extract certain characteristics and match them against similar characteristic of the match. For example - If person in the photo is smiling likely match will be another smiling photo of opposite gender.
So, now you have in a mix known characteristic, preferences, physical characteristics, topics and themes of interest and “DNA” is ready to go on a search for the ideal partner. While conceptually all this is sound, and technologically possible to do – cynic in me is just giggling. Data mining has never been “exact science” so, before I take it any more seriously – I would like to see some match results (preferably from “confusion” matrix.)

No comments:

Post a Comment