Tuesday, February 09, 2016

[igdkftfk] Endorsement graph

We construct three graphs: Person B personally endorses person A as a good potential political leader.  We draw an edge from B to A.

On a second graph, person B believes A is a good judge of character, both for political leadership and for people endorsing political leadership.  Draw an edge from B to A.

Fairly straightforward would be combine these first two graphs to count who has the most indirect endorsements.

On a third graph, we draw an edge from person A to person B if A accepts (or, would accept) B's endorsement.  That is, A trusts B not to be a fraudulent endorser (e.g., a sock puppet) and transitively trusts B not to have accepted fraudulent endorsements.  This too is a directed graph of trust, but a different kind of trust.

All three graphs are public with the nodes containing identifying information.  Third parties attempt to verify them, especially to discover fraudulent endorsements, so exposing those who had poor judgment to accept (perhaps indirectly) fraudulent endorsements.

No comments :