To answer your questions now:
1- This post is only intended to introduce a discussion on influence measurement, so stay tuned for more content heavy posts (1st one already out).
2- I won’t be the person on our team talking “geek”, discussing, and comparing methods to measure online influence. The geek corner (that we’ll have to find a new name for) will be started soon by members of the Traackr team leading our R&D and engineering.
3- My contribution to the topic is to frame the problem we’re talking about here. Math and technology will only provide answers as good the questions we ask, so the framing of these questions is at least as important as the methods we apply to seek answers. This is where I suggest we start.

Hope to see you chime in this conversation!

There is no framework, no ingredient, etc. If you really want to have a geek corner, than start talking geek. Did you use a nonlinear regression, what kind of regression model is it? Is it a support vector regression, neural network for function approximation, or what? You mention machine learning algorithm, so what is the machine learning algorithm called? what optimization procedure did you use to fits that model? What is the loss function you used? did you use square loss, or absolute value deviant loss.

Did you use social network analysis? what defines the edge relationship in the network. How do you construct the network? What are the criteria for creating and edge?

At the very least, I’ve shown step by step from raw data, to how to construct the graph, and then name the algorithm that computes the score.
http://lithosphere.lithium.com/t5/Building-Community-the-Platform/Community-Influencers-Step-by-Step/ba-p/5712

Say something more technical/theoretical and of substance so the techy people can chime in. I didn’t go into the technical detail in my blog either, but my blog is not intended to be a geeks corner. But if you want to talk geek, I’m in! Science and math is like food to me. I happy to discuss that at any level of depth you want, but give me something deeper and scientifically interesting enough to start with, so I can continue.