Information Technology Reference
In-Depth Information
reasoning procedure, in order to determine the importance of every factor and so
its degree of influence on the corresponding assignment. Usually every individual,
in order to conclude to a decision, doesn't take into consideration all the possible
factors but focuses on the most important factors, a procedure that is dependent
on the specific conditions; that means the same expert, in another case, may select
another set of essential factors.
In order to generalize the procedure and produce a generic decision making pro-
cedure, the following approach is introduced. First, the possible factors that may
influence a decision are determined based on bibliographic and general accepted
methodologies, then specific cases are presented to a group of experts, asking them
to select the most important factors for each case and coming to a decision based
on these factors. Thus, for every case, each expert usually selects 3-5 factors, based
on his or her experience, from which decision/ diagnosis is concluded. So for every
factor / concept, we introduce its
importance weight
, which will be used then to
determine its influence to the final decision:
# of experts considering this factor
total number of cases
iw
=
.
(27.3)
Moreover, we introduce a complementary second weight, the “influence to specific
decision”
specific weight
-
sw
, which represents how much the specific factor leads
towards a specific decision / diagnosis. The procedure to calculate the is the fol-
lowing, every expert who considers one factor as important and he takes it into con-
sideration, he is asked to present the degree with which the specific factor leads the
expert to select one decision. Every expert describes the degree of influence of one
factor towards one decision using a linguistic variable, such as “strong influence”,
“medium influence”, “weak influence”, etc.
More specifically, the causal interrelationships from one factor/ concept towards
a decision/ diagnosis concept are declared using the variable
Influence
which is
interpreted as a linguistic variable taking values in the universe
U
.Its
term set
T(influence)
is suggested to comprise nine variables so that to permit to
experts to explicitly describe the degree of influence, actually using nine linguistic
variables, an expert can describe in detail the influence of factor concept towards
decision concept and can discern between different degrees of influence. The nine
variables used here are:
T(influence)
= negatively very strong, negatively strong,
negatively medium, negatively weak, zero, positively weak, positively medium, pos-
itively strong and positively very strong. The corresponding membership functions
for these terms are shown in Figure 27.3 and they are
=[
−
1
,
1
]
μ
nvs
,
μ
ns
,
μ
nm
,
μ
z
,
μ
pw
,
μ
pm
,
μ
ps
,
and
μ
pvs
.
Thus, every expert describes the
specific weight sw
of each interconnection with
a fuzzy linguistic variable from the above mentioned set, which stands for the re-
lationship between the two concepts and determines the grade of causality between
the two concepts. Then, all the proposed linguistic weights for one interconnection
suggested by experts are aggregated using the SUM method and an overall linguistic
weight is produced. The overall linguistic weight with the defuzzification method of
Center Of Gravity (COG) [21], is transformed to a numerical weight
sw
, belonging
