Information Technology Reference
In-Depth Information
the ability of that doctor. For example, John may have an opinion: I think that the doctor is
quite good at his work . In this case we have the belief's content: ' the doctor is quite good
at his work ' and the belief source: ' John '. Considering, in this specific case, the four factors
above described, we have:
the value of the content (doctor is quite good at his work );
the degree of certainty that the trustor has about the fact that John has expressed this opinion
( Iamsure that John told me (thinks) that, etc.);
how good John considers his own belief (when John says: 'I think', he could mean: Iam
sure/ I am quite sure/ I am not so sure and so on);
the credibility of John's opinion (from the trustor's point of view).
The first factor represents a property, a belief and the value of its content (for example,
ability); it is a source's belief that becomes an object of the trustor's mental world. The second
factor represents a trustor's degree of certainty that the source expressed (communicated) that
belief (it is also linked with the trustor's self-trust). The third factor represents an epistemic
evaluation that the source makes on the communicated belief. Finally, the fourth factor repre-
sents a degree of trust in the source's opinion, and it depends on a set of trustor's beliefs about
source's credibility, ability to judge and so on.
The second, third and the fourth factor are not objects of the same level, but rather meta-
beliefs: they represent a modulation of the beliefs. In our networks, this can be better rep-
resented as impact factors. So, in our network we have two main nodes: 'John's belief' and
'Reputation about ability'. The first factor sets the value of the first node. The second, third
and fourth factors set the value of the edge from the first to the second node.
The impact factors are not evaluation beliefs, but rather epistemic ones: they describe the
way to see the other beliefs and their degree of certainty. So, at the level of building belief
sources, evaluation and epistemic factors are separated; from the belief sources level up, in
our FCM representation, they are combined in a unique numerical value. 2
11.4.1 A Note on Self-Trust
Self-trust is a belief that relies on many beliefs, as, in general, trust is: their belief-FCM
can be built in the same manner. As for trustfulness, self-trust is specific of a task or of a
context. Among belief sources there can be, as usual, personal opinions and others' ones - i.e.
reputation.
In self-trust computation there is also a set of motivational factors: self-image, auto-
deceiving, and so on. Since our implementation does not represent motivational factors, at
this moment we are not able to take into account these factors; so we calculate self-trust in the
same way trust is calculated.
2 Even quantitative information (how much I know about) is combined; for example, a low value about the ability
of a doctor can derive from: low evaluation; low confidence in my information sources, little information.
 
Search WWH ::




Custom Search