Information Technology Reference
In-Depth Information
information about
R
SD
and in that case we may even have
R
SD
=
R
SD
=somelow
value (including 0 if there is no information at all about a given disease).
Notice, that the exclusion relation
R
SD
was introduced in [8] to define Conorm-
Cadiag to be able to establish a correspondence between CADIAG and MYCIN-like
systems. CADIAG-like and MYCIN-like are computer assisted medical diagnosis
systems for different applications. CADIAG-like systems are based on fuzzy rules
and an inference procedure - a composition of fuzzy relations - is applied. MYCIN-
like systems use combining functions to calculate the global weights (degrees) of
suggested diagnoses [8, 12].
We intend here to show different types of
S
D
relations, given by an expert
as initial information, and we do not discuss a possibility to substitude
R
SD
−
by, for
R
SD
.
We may introduce another type of relation
R
t
SD
(
example, 1
−
- temporal or timeofmani-
festation. This relation shows to which degree it is true that exhibition of a symptom
s
leads to immediate manifestation of a disease
d
(or a disease
d
is manifested after
some period of time). Notice, that this relation may be also estimated by linguistic
values from Table 25.1.
Thus, it can be seen that relations between symptoms/signs and diseases can
be of different type (occurrence, confirmation, exclusion, temporal, etc.). And in
general, it depends what information is available, what estimations are given by
expert-physicians.
s
i
,
d
j
)
25.3.2
The Patient Information
To use an approximate reasoning mechanism to infer a diagnosis, information about
a patient to whom a diagnosis will be established has to be available. Although med-
ical knowledge concerning
S
D
relationship constitutes one source of imprecision
and uncertainty in the diagnostic process, the knowledge concerning the state of the
patient constitutes another [16].
Information about patients' symptoms/signs is presented in the form of a fuzzy
set
S
p
:
−
, where each element of the fuzzy set
S
p
shows to which degree it
is true, that a patient
p
has symptom/sign
s
i
, or a degree of possibility of the presence
of the symptom, or its severity; symbol
D
p
is used for the final diagnosis for a patient
p
,
D
p
:
Σ
→
[
0
,
1
]
, where each element of the fuzzy set
D
p
shows to which degree it
is true, that a patient
p
has a given disease
d
j
, or a degree of possibility with which
we can attach each relevant diagnostic label to the patient [16].
In the Section 2 we have described criteria that are used for diseases description.
Due to our denotations, these criteria are used for construction of a knowledge-base
of type, e.g.,
R
SD
,
R
SD
,
R
SD
,
R
t
SD
and also these criteria represent symptoms/signs of
a patient to be estimated. Notice, that in our approach we use linguistic scale (Table
25.1) and its simple numerical representation (Table 25.3) for
S
Δ
→
[
0
,
1
]
D
relations and,
correspondingly, the same estimations (linguistic and numerical) are assumed for a
degree of truth, that a patient
p
has symptom
s
i
'. Additionally note, that, although
−