How is the clustering analysis going to work?

Because the current hard coded list of URLs is a start, but it's not really a scalable solution to the issue.

However, from what I can see in this file:

https://github.com/jacquerie/stop-the-bullshit/blob/master/d...

It just seems like it's going to compare examples of articles included in the source files, as found here:

https://github.com/jacquerie/stop-the-bullshit/tree/master/d...

So how is it going to detect the difference between a real or fake piece from this?