Information Technology Reference
In-Depth Information
The
importance weight
,
iw
, values for these factors were calculated on incidence %
of arrivals in an ED and refers to patients over 65 years of age.
It is concluded that the CFCM-ESI will consist of these total 22 Factor Concepts
and thus, based on them, a possible triage Decision will be assigned. More accu-
rately each patient is assigned one of the 5 ESI levels, therefore the CFCM-MDSS
will include 5 Decision Concepts (DC), each one for every ESI level:
Table 27.2
Decision Concepts
Decision concept (DC) Physical meaning
DC1
ESI Level 1 (ESI1)
DC2
ESI Level 2 (ESI2)
DC3
ESI Level 3 (ESI3)
DC4
ESI Level 4 (ESI4)
DC5
ESI Level 5 (ESI5)
After determining the concepts of the Fuzzy Cognitive Map, the most important
issue is the assignment of the influence among concepts, which is the second step
of FCM development [45]. The FCM development procedure is very important
since this model is then used for decision making and diagnosis. Here, we further
apply the designing methodology presented in section 27.4 in order to assign weight
values between the Factor Concepts (FC) and the Decision Concepts (DC).
This designing methodology utilizes the data used and provided in the study of
Garbez et al., [12], where 334 cases of patients were examined and 18 experts as-
signed them to ESI levels. According to this designing methodology, the first stage
is the assignment of the
importance weight
, at every concept using equation 27.3,
which is depicted in Table 27.1. Then the
specific weight
,
sw
, representing the in-
fluence from a Factor Concept FC to a Decision Concept DC must be determined.
Subsequently equation 27.4 is applied in order to calculate the weight from Factor
Concept to Decision Concepts. Here, for this case, in order to calculate the overall
weight from FCs to DCs, a simplified version of equation 27.4 is used along with
the normalization to 1, where
l
1
=
1and
l
1
=
0
.
5. Thus:
w
ji
=
sgn
(
sw
)(
iw
+
0
.
5
∗|
sw
|
)
(27.5)
The overall weight after the normalization to 1, is then fuzzified according to the
membership functions of Fig 27.3. The weights from FCs to DCs are depicted in
Table 27.3 which are used to produce the CFCM-ESI illustrated in Figure 27.4.
At this stage of the research we only assign weights from Factor Concepts to
Decision Concepts, but the FCM capabilities permit us to introduce weights among
the Factor Concept themselves, that create a more accurate but too complex model,
which is part of ongoing research.
