Information Technology Reference
In-Depth Information
based on a dynamical model of preference aggregation. The fundamental problem
becomes the construction of an appropriate consensus measure. In this paper, as-
suming that the opinions of medical specialists are represented by numerical fuzzy
preferences, we develop a soft consensus model combining a measure of collective
dissensus with an inertial mechanism of opinion changing aversion.
The structure of the paper is as follows. In section 2 a background explanation
of the diagnostic framework in which our consensus module is embedded will be
given. Section 3 focuses on a short historical overview of the main contributions
in the area of group decision making and consensus modeling under fuzziness. In
section 4 the model describing the consensual dynamics is presented.
21.2
A Diagnostic Frame
Sadegh-Zadeh in [25] wrote: “Medical diagnostics is a spatio-temporal network of
collective action, a node of which accommodates an individual patient or a group of
patients ... medical diagnosis is a social construct because the process of diagnostics
that produces the diagnosis is itself socially shaped”
And then introduced the following 10-tuple diagnostic frame
•
P
, set of patients
•
PO
, a population that
P
is a subset of
•
E
, set of diagnosticians (experts)
•
D
and
Δ
, sets of categorical and conjectural abnormality statements on
P
•
G
, set of goals the diagnosticians pursue
•
A
, set of actions available in pursuing goals
•
KB
, knowledge-base used by diagnosticians
•
M
, set of models (KBMS, optimization algorithms, reasoning processes, etc)
that guides the diagnosticians from
(
D
,
KB
)
to
Δ
.
•
T
, the time period
In this frame, a central role is played by the collective medical decision process
where experts from different medical fields, mostly in a “same place same time”
meeting, sharing text and images, have to discuss and to agree about a common
diagnosis and treatment.
Therefore, since evidence suggests that diagnoses made through collaboration
achieve a higher performance then ones made by an individual clinician, addressing
the problem involving several different specialists aiming at finding a consensual
diagnosis is an approach becoming in more widespread use day by day. Accord-
ingly, the architecture of the GDSS supporting the collaborative work of the group
of clinical experts can be summarized as in Fig.21.1.
From the point of view of group decision theory, reaching a consensual diagnosis
means updating opinions with respect to a given set of alternative diagnoses where
these opinions are usually represented by preference relations. The uncertainty gen-
erated by imperfect, imprecise, information is a vital part of the diagnosis problem
itself which should be combined with the vagueness in the clinicians' way of think-
ing (fuzzy mode of thinking). Consequently, the use of fuzzy relations in medical
