Mining Matrix Pattern from Mobile Users - Distributed Artificial Intelligence, Agent Technology, and Collaborative Applications - page 247

Step6:CalculateRelationshipConidenceforEachSingleLayer
Figure 13 shows that after significant confidence is calculated, the time has come to calculate the rela-
tionship in between the confidence for each single layer. The relationship in between each static node
in each layer must involve two and only two static nodes due to the nature of relationship. The nature of
the relationship is the association behaviour, or association characteristics in between two entities, and
in this case, the static nodes. The formula to calculate relationship confidence is: relationship confidence.
= (static node confidence 1 + static node confidence 2) / 2 .
Step7:SelectRelationshipConidenceThresholdforEachSingleLayer
After the relationship confidences are calculated, those relationship confidences that are insignificant
must be screened out. At this level, the decision maker can be quite confident that the confidence re-
lationships obtained are highly significant and the use of threshold values to screen out relationship
confidence may be limited to confidence threshold + 10%. After the relationship confidence threshold
is determined, all relationships with relationship confidence lesser than the relationship confidence
threshold will be marked insignificant, while all relationship confidence greater than the threshold will
be marked as significant.
Step8:CalculateMulti-LayerRelationshipConidence
After both the confidence and relationship confidence have been calculated, it is now necessary to
calculate the multi-layer relationship confidence. This is the core of the multi-layer matrix pattern in
which it provides a systematic way for decision makers to select multiple layers of the matrix depend-
ing on the areas of logical domains that the decision makers wish to take account into. The different
selection of logical domain areas leading to different results will further allow the decision maker to
compare the results of the tests.
Figure 14. Algorithm for calculation of multi-layer matrix confidence
Confidence Calc_Relationship_Conf (Conf1, Conf2)
C1OK = Conf1.validate();
C2OK = Conf2.validate();
If (C1OK && C2OK) {
Confidence = Average(Conf1, Conf2);
If (Confidence < Confidence_Threshold) {
Confidence = 0;
Return Confidence;
} Else {
Conf1.Location.Mark();
Conf2.Location.Mark();
Conf_Table3D.Set(Conf1, Conf2, Confidence);
}
}
}