Imprecise Communication and Pattern Matching in Medical Diagnosis(Fuzzy Logics in Biomedical Systems)

(1)

Biomedical Fuzzy Systems Association

NII-Electronic Library Service

BiomedicalFuzzySystemsAssociation

Biomed

Fuzzy

Sys.

_Bul1,

_Vol.1,

No.2,1991

Imprecise

Communication

and

Pattern

Matching

in

Medical

Diagnosis

Ronald R. Yager

Machine Inteliigenee Institute Iona

College

New

Rochelle,

NY

10801

USA

'

Medtcal

diagnosis

is a complex and many

_leveled

_proeess. _At

a fundamental level it ean be seen as a _pattern matching _problem

where one is trytng to mateh symptoms _w!th

_disorders.

rmprecision and uncertainty _pervade _thls _matchlng

proeess. Among

the most notable sources _are _vague _and _unspecific _definitions

of

diseases and associated symptoms.

_In

_many _eases _only

a

statistical correlation extsts between _symptoms _and

_dtsorders.

a

further compounding factor is that

decislons

and actions must

be

taken in the face of

tncomplete

informatton.

_Thts

sometimes

requtres the use of default rules, such as the easiest first

rule.

Another

level

of the process tnvolves a

dialogue

between

_a

patient and a doctor.

This

coinmunlcation

_tnterfa

¢e between the

doctor and the

_patient

is

_an _essenttal

component ef _the

information

gathering _process. _As

_tn

_any _other _{tnterpersenal}

relatlenshtp tt is pervaded by various

ktnds

of uncertainty and

Iack of

frankness.

Information

_provided

_by

_patients _is _usually

of an ±mpressionlsttc _nature.

_.Pattents

_don't _always

provide ell

the relevant

inofrmation,

sometiMes

from

_ignoran

¢ e and sometimes

intenttonally when they

feel

that

tt

will cenfltct wlth some

other agenda of thetrs.

In this talk we sha)l discuss a number of _techntques, _based

on

tuzzy

set theory,

for

managing _the _{uncertainttes}

_in

_the dlagnostlc envtronment.