Jon Marra, the director of data science at eHarmony, was recently interviewed about the processes the US dating site uses to connect its members.
Marra said in the interview, with Data Skeptic, he believes “the notion of love is potentially quantifiable,” by using a combination of both machine learning and data science.
Although he recognises that a lot of factors come into play when matching people together, Marra said eHarmony uses a three step process to match their members, which focuses on users’ compatibility and affinity, as well as using a graph flow optimisation to make sure everyone has an equal chance to find their best match.
Marra says the process seeks to go further than matching people on just their likes and dislikes, looking into other factors such as fiducial markers (features), amongst others, in order to ensure people get the best matches possible.
To hear the full interview please click here.