Page 49 - Trust in ICT 2017

Trust in ICT 1

level, both credentials and reputation involve the transfer of trust from one entity to another, but each
approach has its own unique problems which have motivated much of the existing work in trust.

A trust decision can be a transitive process, where trusting one piece of information or information source
requires trusting another associated source. For example, one might trust a book and its author because of
the publisher, and the publisher may be trusted only because of the recommendation of a friend. Winslett’s
work [14] in policy-based trust uses (or refers to) “credential chains” (the issuer of one credential is the
subject of another), the majority of transitive trust computation has been focused on using reputation. A key
recent example of this approach is Golbeck and Hendler [15] [16], which describe how trust is computed for
the application TrustMail. Reputation is defined as a measure of trust, and each entity maintains reputation
information on other entities, thus creating a “web”, that is called a web of trust.

6.2.3.2 Trust Ontology
It is needed to use of a knowledge base for storing trust models and trust related context specific data that
does not alter the calculations or use of trust related information, such as reputation (entity opinions). The
knowledge base should clarify how information is stored and accessed and ontology is one of the prospective
solution. For example, a trust network can be seen as a structure capturing metadata on a web of individuals
with annotations about their trustworthiness. Considering social network as our context, a trust network can
be seen as an overlay above the social network that carries trust annotations of the metadata based on the
social network, such as user profiles and information.

Social networks are gaining increasing popularity on the web while semantic web and its related
technologies, are trying to bring social networks to their next level. Social networks are using the semantic
web technologies to merge and integrate the social networking user profiles and information. Such efforts
are paving the path toward semantic web-driven social ecosystems. Merging and integrating social
networking data and information can be of business value and use to web service consumers as well as to
web service providers of social systems and networks. Ontologies, at the core of semantic-web driven
technologies lead the evolution of social systems on the web. Describing trust relations and their
subcomponents using ontologies, creates a methodology and mechanism in order to efficiently design and
engineer trust networks.

“Structure of a given system is the way by which their components interconnect with no changes in their
organization”. Determining the structure of a society of agents on a trust network structure within a semantic
social system, can help us determine the organizational structure of a system. Having this capability, an
organization’s certain factors such as flexibility, change capacity, etc., can be determined.
The work by Golbeck and Hendler uses ontologies to express trust and reputation information, which then
allows a quantification of trust for use in algorithms to make a trust decision about any two entities. The
quantification of this trust and associated algorithms are called trust metrics. Given an existing quantification
of trust, approaches exist to transfer that trust to other entities, which may not have been evaluated for
trust. One area of research assumes we are given a web of trust, where a link between two entities mean a
trust decision has been made and the value of that trust is known. How trust decisions are made do not
matter, as long as the resulting trust values can be quantified. If there is no link between a pair of entities, it
means no trust decision has yet been made. This is the case in which trust transitivity can be applied, a
simplified example being if A trusts B and B trusts C, then A trusts C. Building on work in reputation
management (described earlier as empowering individual agents to make trust decisions instead of a single,
central authority making decisions for them), multiple researchers are exploring ways to transfer trust within
a web of trust.
6.2.4 Reputation and Trust Analytic

Reputation is third-party information and is considered as both social product and social process. It is a social
product because it is produced by opinions of entities; on the other hand, reputation is as an information
flow influencing in the social IoT. Reputation should not to be confused with trust but partially affects the
trust. There are several well-known reputation systems in the context of e-commerce systems, such as eBay
[17] and Internet-based systems such as Keynote [18]. These systems use a centralized trust authority to
maintain the reputation and feedbacks. There are also some distributed approaches for reputation

41