what-when-how
In Depth Tutorials and Information
A story can be classified as fact, rumor, or questionable.
≥
1
ϕ
⎧
fact Trust
(
(
)
Trust
)
sender
rcvr
⎪
≤ −
1
ϕ
story
rumor
(
Trust
(
)
Trust
)
⎨
(11.35)
sender
rcvr
⎪
⎪
questionable el
(
se
)
⎩
where
φ
≤ 1
is a threshold value set by the receiver,
Trust
sender
and
Trust
revr
are the
trust of the story sender and receiver, respectively.
If a story is a fact, it will be accepted by the local node, included in the local
aggregate, and disseminated to the neighboring nodes. he story classiied as a
rumor will be discarded. For the questionable story the receiver will ask for the
additional information by sending a review request to a random set of nodes [27].
here is a rating process after acceptance called
postreviewing of stories
where an
augmented Bayesian trust estimation [28] process can be used to compute the
trust values which will be associated with the story senders according to the rating
obtained from the post review.
11.4.10 TrustDavis
In Reference 26 it proposes a trust model called TrustDavis
based on the insurance
mechanism against malicious behaviors in the transaction. he main contributions
of TrustDavis are [26]:
◾
Honest nodes can limit the damage caused by collusion of malicious nodes.
◾
he multiple identities of malicious nodes can be limited in the system.
◾
he nodes incline to rate other nodes accurately.
◾
TrustDavis is distributed.
In TrustDavis, nodes initially publish references, which is an acceptance of limited
liability. hen the social relationships among the nodes are established. To join
in the network the newcomer must provide the security deposits to a trustworthy
node for the reference. his rule facilitates to limit the multiple identities of mali-
cious nodes. Based on the insurance concept, the payment for reference and the
payback in case the transaction fails can be easily calculated. Besides, TrustDavis
adopts the strategy for the false claim scenario.
11.5 FutureResearch
he trust models we mentioned above are mostly originated from the social net-
works that constitute the top layer in the STNs. Although generally these trust
models may adapt to the environment of the top layer of the STNs, some problems
