Information Technology Reference
In-Depth Information
to the interval
. A similar approach was initially presented for the description
of the development of FCM model in [45].
Then the overall weight describing the influence from one factor concept towards
a decision concept is calculated using the form:
[
−
1
,
1
]
w
ji
=
sgn
(
sw
)(
l
1
∗
iw
+
l
2
∗|
sw
|
)
(27.4)
where the two parameters
l
1
,
l
2
are introduced to represent the participation of the
importance weight iw
and the
specific weight sw
, on the overall weight describing
the influence of every factor concept towards the decision/diagnosis concept. It is
mentioned that the value of
w
ii
has to be normalized in the interval
[
−
1
,
1
]
,where
the weight takes values.
Fig. 27.3
Membership functions of the linguistic variable Influence
27.4.1
Developing Fuzzy Cognitive Maps Model for the 5-Level
ESI Triage System
The described methodology on developing FCMs is implemented for the case of
constructing a medical decision making system for the 5-level Emergency Severity
Index (ESI) triage system [10, 16, 46, 53]. The ESI describes the main factors
and based on them provides a standardized algorithm for the triage process using a
systematic approach and utilizes both intuitive and analytical approaches to clinical
decision making.
Based on the literature as presented in sections 27.3.2 and 27.3.3, as well as a
study of 18 triage nurses [12], a series of factor concepts were concluded to be
part of the Competitive Fuzzy Cognitive Map Medical Decision Support System
for ESI (CFCM-ESI) for triage. Twenty-two (22) factors are selected that represent
the potential concepts of the Fuzzy Cognitive Map decision model, but they do not
all have the same importance in order to infer an assignment on the severity of the
patient based on the 5-level triage system.
In the research of Garbez et al. [12], only ESI level 2 or level 3 were stud-
ied, where triage nurses were asked to select 3 to 4 factors that they rated as
