TrustGraph

https://github.com/AdamColton/trustGraph

TrustGraph

Nodes represent entities (users, identities or processes) and edges represent trust. Trust is directional and weighted. The trust graph is a way of asking if a user can trust some proposition. Generally this is going to be along the lines of asking if a piece of software is secure or a set of data is accurate. Each node can provide an endorsement at a minimum confidence level.

An important detail is that user rarely will calculate the exact level of trust they have in something. Instead, they will be looking for a minimum trust level. They will then cascade out until they either reach that level or conclude that the rest of the graph cannot provide the assurance they need. This is important to reduce the complexity of the calculation.

Uses EigenTrust. http://nlp.stanford.edu/pubs/eigentrust.pdf