Reputation assessment in collaborative environments.

The popularity of open collaboration platforms is strongly related to the popularity of Internet: the growing of the latter (in technology and users) is a spring to the former. With the advent of Web 2.0, not only the Internet users became from passive receiver of published content to active producer of content, but also active reviewers and editors of content. With the increase of popularity of these platforms, some new interesting problems arise related on how to choose the best one, how to choose the collaborators and how evaluate the quality of the final work. This evolution has brought much benefit to the Internet community, especially related to the availability of free content, but also gave rise to the problem of how much this content, or these people, may be trusted. The purpose of this thesis is to present different reputation systems suitable for collaborative environments; to show that we must use very different techniques to obtain the best from the data we are dealing with and, eventually, to compare reputations systems and recommender systems and show that, under some strict circumstances, they become similar enough and we can just make minor adjustment to one to obtain the other.