Online communities are becoming important in all aspects of
our lives, including businesses, health care, personal
relations and so on. Very successful examples of online communities include
social networks, such as Facebook, Twitter, etc., and other systems based on
user reviews and their interactions, such as eBay, Amazon, Yelp, Epinions, etc. Online communities generate large amounts of
data that contain useful information mixed with noise. The noise can be random
(for example some irrelevant reviews) or on purposely injected (various
attacks, such as fake information). In our real life’s social interactions we
rely heavily on trust, similarly, in online communities, we can use trust to
filter out all kinds of noises and extract the useful information.
Therefore, we have developed develop a modular and general Trust Management
Framework, to capture the specificity of given communities and apply common rules
of trust, both based on social science results on human trust. We have applied
our trust framework successfully on Twitter and Epinions.
Furthermore, we are using it in multi-party decision making, as well as in
distributed computing environments.