The old way.
It was one dimensional using only online dater self-populated data fields.
f(a, b, g)
The "f" function is just user "self populated data" : Subscriber's city, body type, hobby preferences, favorite activities, etc. This set of "lies" was matched against another set of 'self populated' inputs.
The new way.
Use third party matching analysis (TPMA) to battle online dating fraud. The additional epsilon axis is for third party data. Third party data is harder to manipulate, not user self populated and less biased.
Epsilon is "third party data" for Online Dating and Online Personals
The New Way-
Your matches get an additional dimension - The epsilon function