Homeopathic Drug Selection Using Intuitionistic Fuzzy Sets

Athar Kharal

Centre for Advanced Studies in Pure and Applied Mathematics,

Bahauddin Zakariya University, Multan, PAKISTAN
Email: atharkharal@gmail.com
Abstract. Using intuitionistic fuzzy set theory, Sanchezs approach for med-
ical diagnosis has been applied to the problem of selection of single remedy from
homeopathic repertorization. Two types of compositions of IFRs and three types of
selection indices have been discussed. We also propose a new repertory exploiting
the benits of sof-intelligence.
Keywords: Intutionistic fuzzy set; Intutionistic fuzzy relation; Homeopathic drug
selection; Automated repertorization; Automated decision making; Homeopathic
software
1. Introduction
The notion of intuitionistic fuzzy sets (IFSs) was introduced by Atanassov [1] in 1986. It
is suciently generalized a notion to include both fuzzy sets [14] and vague sets [6] as
special cases. Fuzzy sets are IFSs but the converse is not necessarily true [1]. Whereas
the notion of vague sets has been proven to be the same as that of IFSs [5].
IFSs have been found to be very useful in diversely applied areas of science and tech-
nology. In fact there are situations where IFS theory is more appropriate to deal with
[4]. To name only a few areas of applications, IFSs have been applied to logic program-
ming ([2] [3]), medical diagnosis, decision making problems [13] and microelectronic fault
analysis [11].
Homeopathy has often been called the third-most commonly used system of healing
on the globe. For that reason alone it deserves serious attention from the modern scien-
tic community. In homeopathic practice, use of repertory is inevitable for a successful
application. This very tedious process of repertorization of a patients case has now been
facilitated by the commercially available software like CARA and RADAR. But such
software, instead of suggesting a single remedy, leave the practitioner with about ten to
fteen drugs, from which she has to decide for the single one. This, in fact, is the intrinsic
shortcoming of the very nature of repertorial work, as alluded by Kent himself:
"..... method [of repertorizing] is the mechanical method ..... There is
an artistic method that omits the mechanical, and is better, but all are not
prepared to use it. ......... Symptoms must be judged as to their value as
characteristics in relation to the patient ........... They must be valued in
proportion as they relate to the patient ..... Any remedy correctly worked
out, ..... should be perceived to agree with, and to t, the patient ...." [7]
Hence a need is there, once the mechanical repertorization is achieved, to include some
module of automated decision in software systems for selecting the nal single homeopathic

Phone: +92 333 6261309

1
Homeopathic Drug Selection Using Intuitionistic Fuzzy Sets 2
remedy i.e. similum. It is noteworthy that such an automatization of drug selection does
not replace the homeopathic practitioner on two counts: rst that the gradation and
choice of symptoms lies in her hands and secondly, the Intuitionistic Homeopathic Knowl-
edge (to be described latter) is elicited from the experience and provings of homeopathic
practitioners and researchers. This automatization is an attempt to capture and stan-
darize the process of nding the remedy which is most similar to the disease picture of a
patient. This does not, by any means, replace the fundamental Hahnemannian philosophy
of nding the similium.
In the present paper, we modify Sanchezs method of medical diagnosis ([9],[10]) using
IFS theory. After a detailed description we give an illustrative example as an application of
the method. In the sequel we discuss two types of compositions and three selection indices
besides proposing a mechanism for accumulating systematic homeopathic knowledge and
hence a new repertory.
2. Preliminaries
Denition 2.1. [1] Let A be a nonempty xed set. An intuitionistic fuzzy set is an
object of the form = r, j(r), (r) : r A, where j and are degrees of membership
and nonmembership of each r A, respectively, and 0 _ j(r) +i(r) _ 1 for each r A.
For the sake of brevity, we shall use the notation = j
A
,
A
instead of =
r, j
A
,
A
[ r A .
The number
A
(r) = 1 (j
A
(r) +
A
(r)) is called hesitation of IFS A, which may
relate to either membership value or non-membership value or both.
Denition 2.2. [1] Let A be nonempty set and , 1 IFSs in A, given as =< j
A
,
A
,
1 =< j
B
,
B
. Then
1. _ 1 i j
A
(r) _ j
B
(r) and
A
(r) _
B
(r) for all r A;
2. = 1 i _ 1 and 1 _ ;
3. = r,
A
(r), j
A
(r) : r A ;
4.
T
1 = r, j
A
. j
B
,
A
.
B
: r A ;
5.
S
1 = r, j
A
(r) . j
B
(r) ,
A
(r) .
B
(r) : r A .
Denition 2.3. [5] An intuitionistic fuzzy relation (IFR, for short) 1 from a set A to
another set 1 is an IFS in A 1 i.e.
1 = (r, j) , j
R
(r, j) ,
R
(r, j) : r A, j 1 ,
where j
R
: A 1 [0, 1] and
R
: A 1 [0, 1] satisfy the condition
(\(r, j) A 1 ) (j
R
(r, j) +
R
(r, j) _ 1) .
An IFR 1 from A to 1 is denoted by 1
XY
.
Homeopathic Drug Selection Using Intuitionistic Fuzzy Sets 3
Denition 2.4. Let Q
XY
and 1
Y Z
be two IFRs. The max-average composition 1oQ
is the intuitionistic fuzzy relation from A to 7, dened by the membership and non-
membership functions, respectively, as:
j
RQ
(r, .) =
_
y
[
1
2

j
Q
(r, j) +j
R
(j, .)

] and
RQ
(r, .) =
^
y
[
1
2

Q
(r, j) .
R
(j, .)

]
\(r, .) A 7 and \j 1.
Denition 2.5. [4] Let Q
XY
and 1
Y Z
be two IFRs. The max-min-max composition
1oQ is the intuitionistic fuzzy relation from A to 7, dened by the membership and the
non-membership functions, respectively, as:
j
RQ
(r, .) =
_
y
[j
Q
(r, j) . j
R
(j, .)] and
RQ
(r, .) =
^
y
[
Q
(r, j) .
R
(j, .)]
\(r, .) A 7 and \j 1.
3. Sanchezs Scheme in terms of IFSs
By using IFSs we rst generalize and adapt Sanchezs scheme for medical diagnosis ([9],
[10]) to the problem of selection of homeopathic similum. In a given homeopathic case,
suppose o is the set of symptoms, 1 set of drugs, and 1 the set of patients. Analogous to
Sanchezs notion of "Medical Knowledge" we dene Intuitionistic Homeopathic Knowledge
(IHK for short) as an intuitionistic fuzzy relation 1 from the set of symptoms o to the
set of drugs 1 (i.e., on o 1) which reveals the degree of association and the degree of
non-association between symptoms and drugs.
3.1. Method. The methodology of intuitionistic fuzzy drug selection comprises fol-
lowing main steps:
1. Determination of symptoms of the patient and their associations/non-association to
her by routine case-taking. Thus mathematically, a patient is an IFS, say , on the
set of symptoms o. This IFS would help to construct an IFR C
PS
, relating patient
to symptoms.
It is customary to underline symptoms reported by patients with more underlinings
indicating stronger symptoms. It must be noted that assigning association and non-
association degrees to each symptom is a generalization of the underlining practice.
For example if a practitioner wants to use only traditional underlinings, then he
may x association degree of one underline as 0.3, double underline as 0.6 and three
underlines as 1. But the provision of discerning more grades of symptoms than the
traditional three, allows for a ner analysis and thus gives an edge to this method
over the classical one.
Being part of a generalized method association and non-association degrees
may be interpreted variously. For example rst number of a pair is the degree to
which a patient thinks the symptom to be important and the second is the degree
of importance attached to it by the practitioner. Patient may attach a great value
to his aching knee but for practitioner it is just a particular physical symptom and
Homeopathic Drug Selection Using Intuitionistic Fuzzy Sets 4
not a generality or modality. The generality of method includes the classical case
as a special one as follows: it is clear that
j(r) + (r) _ 1, \r A
allows us to select non-association (r) as
(r) _ 1 j(r) .
This implies that if only the traditional analysis is to be adhered to, one should use
the mechanical formula for the non-association i.e.
(r) = 1 j(r) . (1)
Clearly, in this case one deals with a subclass of IFSs i.e. fuzzy sets and, in fact,
would be under-utilizing this method. While using (1), this should also be kept
in mind that it is bit debatable if an association of (say) 33% marks mechanically
implies a two-third non-association.
2. Codication of homeopathic knowledge as an intuitionistic fuzzy relation. It would
be achieved by the help of an expert homeopath or some other dependable source
of homeopathic knowledge, for example as the one proposed in [8]. If the 3 or 5
grade system used by dierent repertories is required to be carried along, itmay be
handled as was done in case of patient-symptom underlinings in step 1 above. This
IFR from o to 1 of homeopathic knowledge is denoted as 1
SD
.
3. Determination of patient-drug relational strength through composition of intuition-
istic fuzzy relations. The max-min-max or max-average composition 1
PD
of C
PS
with the IFR 1
SD
denoted by 1 = 1oC (may interestingly be read as, 1esult
= 1nowledge applied to Case) signies the patient-drug relation as an IFS on 1
with the membership and non-membership functions, respectively, given by
j
R
(d) =
_
s2S
[j
A
(:) . j
R
(:, d)] and
R
(d) =
^
s2S
[
A
(:) .
R
(:, d)]
\d 1. The choice of composition method lies with the user. If users knowledge
of homeopathic case is to be considered, then Denition 2.4 should be employed. If
one decides to depend upon the knowledge of expert solely, then choice of Denition
2.5 is logically more correct.
4. Computation of selection index o
K
for nal decision. Any one of the following
selection indices may be used:
o
1
K
= j
K

K
.
K
o
2
K
= j
K

K
+
K
2

o
3
K
=

j
K
+
K
2

K
o
2
K
is more discriminatory as compared to o
1
K
. Whereas o
3
K
is rmer on giving
judgement and seldom comes up with a tie decision. Hence rare, if ever, use of
hesitations is made in o
3
K
.
Homeopathic Drug Selection Using Intuitionistic Fuzzy Sets 5
5. In case of a tie decision in step 4, hesitations are called in to get nal judgement
and the drug with least hesitation is selected.
3.2. Generalized Method. This method can easily be extended to a nite number
of patients as follows: Let there be : patients j
i
, i = 1, 2, ..., :, in a hospital. Thus j
i
1.
Let 1
SD
be an IFR on o 1 and construct an IFR C
PS
from the set of patients to the
set of symptoms o. Clearly, the composition 1 of IFRs 1 and C (1 = 1oC) describes
the state of patients j
i
in terms of the diagnosis as an IFR from 1 to 1 given by the
membership and non-membership functions, respectively, as:
j
R
(j
i
, d) =
_
s2S
[j
C
(j
i
, :) . j
K
(:, d)] and
R
(j
i
, d) =
^
s2S
[
C
(j
i
, :) .
K
(:, d)]
or if the max-aveage composition is used, then
j
R
(j
i
, d) =
_
s2S
[
1
2
(j
C
(j
i
, :) +j
R
(:, d))] and
R
(r, .) =
^
s2S
[
1
2
(
C
(j
i
, :) +
K
(:, d))]
\j
i
1 and d 1.
3.3. Accumulation of Homeopathic Knowledge. For a given 1 and C, the rela-
tion 1 = 1oC can be computed. Conversely, from the knowledge of C and 1, one may
also compute a rened version of the IFR 1 (knowledge) such that the following hold
true:
1. selection index is greatest, and
2. the equality 1 = 1oC is retained.
This rened version of 1 will be a more signicant IFR attributing higher degrees
of association and lower degrees of non-association of symptoms as well as lower degrees
of hesitation to the drugs. Thus this is one of the many possible approaches towards
Accumulated Intuitionistic Homeopathic Knowledge. From a rened version of 1, one
may infer drugs from symptoms in the sense of a paired value, one being the degree
of association and other the degree of non-association. In case, the homeopath is not
satised with the results, 1 is modied. This improvement of 1 over a period of time
would establish a rubric in a repertory which would be the compilation of all such 1
0
:. A
computer-based knowledge acquisition module can be used for this purpose. This module
would use the method presented in this work as a means to accumulate homeopathic
knowledge, more precisely, the symptom-drug associations and non-associations, from
dierent practitioners. A brief ow-chart for this module is given below:
It is important to note that this ow chart and specially the loop based upon expert
satisfaction is not part of this IFS method presented in this paper, instead it is the
mechanism to elicit IHK which if gathered and compiled would yield a new repertory in
which entries, for example, may look as follows:
Heat, foot, soles, uncovers them : Am-c
(:4;:5)
, Calc-c
(:8;:2)
, Cham
(:8;:1)
, Con
(:2;:2)
,
Lyc
(:8;:1)
, Nux-v
(:5;:3)
, Puls
(:6;:3)
, Sep
(:8;:1)
,
Sil
(:5;:2)
, Sulph
(:5;:1)
Homeopathic Drug Selection Using Intuitionistic Fuzzy Sets 6
Modify K
Start
Find K for which SK is greatest
retaining the relation R=KoC
Is K
unique ?
Find K for which
hesitation is least
Compute R=KoC
OK
Input case
record
Retrieve K
Determine P, S and CPS
Is the expert
satisfied with the
results ?
yes
no
yes
no
Stop
4. Illustrative Example
To see an application of the method, in the following we give an example (patients
personal details have been omitted due to medical privacy laws) :
Suppose the patient is John whose symptoms were recorded by routine case-taking
practice. A pair of values was attached to each symptom: rst value showing the strength
of association of that symptom with the patient and the other showing non-association
as perceived by the practitioner. The symptoms, practitioner decided to include, looked
something like this:
John had under gone an operation of stomach and since then had been
suering from vomiting and retching (.3, .3) . He has a life long complaint
of coryza (.6, .2) . He becomes angry once his opinion is refuted or disagreed
(.4, .1) . He avoids a shower (.7, .2) . He complained of soles burning specially
in the bed (.8, .1). His stool is rst hard and then liquid (.9, 0) .
The case was repertorized by using the repertories of both Kent and Boericke. Fol-
lowing rubrics were considered:
:
1
= [kent ] [Mind] Anger, irascibility : Contradiction, from
:
2
= [kent ] [Generalities] Bathing:Dread of
Homeopathic Drug Selection Using Intuitionistic Fuzzy Sets 7
:
3
= [kent ] [Extremities] Heat: Foot: Sole: Uncovers them
:
4
= [boericke ] [Stomach] Vomiting & retching: Cause: Post-operative (laparotomy)
:
5
= [boericke ] [Nose] Coryza: Coryza, with chronic tendency
:
6
= [kent ] [Stool] Hard: First: Then uid
Normal repertorization showed following ten drugs to be the most indicated ones:
Cal-c, Puls, Lyc, Am-c, Nux-v, Si l, Sep, Sulph, Cham, Con
Now we proceed to nd out best drug based upon IHK. These rubrics/symptoms have
been named arbitrarily :
1
, ..., :
6
to facilitate their writing as a set. Clearly, 1 = Jo/:
and the set of symptoms o = :
1
, :
2
, :
3
, :
4
, :
5
, :
6
.
The IFR C
PS
is given as:
C
PS
:
1
:
2
:
3
:
4
:
5
:
6
John (.4, .1) (.7, .2) (.8, .1) (.3, .3) (.6, .2) (.9, 0)
The IFR 1
SD
(IHK) is given as follows:
1
SD
Calc-c Puls Lyc Am-c Nux-v Sil Sep Sulph Cham Con
:
1
(.8, .2) (.6, .3) (.8, .1) (.4, .5) (.5, .3) (.5, .2) (.8, .1) (.5, .1) (.8, .1) (.2, .2)
:
2
(.4, .1) (.5, .3) (.8, .1) (.3, .5) (.9, 0) (.7, .1) (.6, .3) (.4, .2) (.2, .2) (.5, .3)
:
3
(.6, .3) (.7, .1) (.4, .6) (.5, .2) (.5, .3) (.9, 0) (.3, .2) (.6, .2) (.3, .2) (.5, .4)
:
4
(.7, .2) (.6, .1) (.8, 0) (.7, .3) (.6, .1) (.8, .1) (.3, .1) (.6, .3) (.4, .3) (.2, .7)
:
5
(.8, .1) (.5, .1) (.3, .6) (.7, .2) (.5, .5) (.3, .5) (.4, .4) (.5, .1) (.1, .2) (.4, .3)
:
6
(.7, .2) (.7, .2) (.5, .3) (.2, .6) (.5, .3) (.7, .1) (.4, .4) (.7, .1) (.5, 0) (.7, .2)
Using max-average composition, 1 = 1oC is computed as:
1
PD
Calc-c Puls Lyc Am-c Nux-v Sil Sep Sulph Cham Con
John (.8, .1) (.8, .1) (.8, .1) (.7, .2) (.8, .1) (.9, .1) (.7, .1) (.8, .1) (.7, 0) (.8, .1)
In the following table o
i
K
; i = 1, 2, 3 are given:
Calc-c Puls Lyc Am-c Nux-v Sil Sep Sulph Cham Con
o
1
K
.8 .6 .8 .4 .4 .4 .8 .5 .8 .1
o
2
K
.7 .4 .7 .1 .3 .3 .7 .3 .7 .2
o
3
K
.5 .4 .4 .4 .2 .1 .4 .1 .4 .4
o
1
K
and o
2
K
do not produce any substantial dierence about the decision, accept some
numerical dierence in their index values. Both these indices render Calc-c, Lyco, Sep
and Cham as equal candidates as o
1
K
= 0.8 = o
2
K
for all of them. Hence we calculate the
hesitations of each drug, based upon 1
PD
, as follows:
Homeopathic Drug Selection Using Intuitionistic Fuzzy Sets 8
Calc-c Puls Lyc Am-c Nux-v Sil Sep Sulph Cham Con
.1 .1 .2 .2 .1 .1 .3 .2 .3 .1
Since least hesitation among the four candidate drugs is that of Calc-c, hence this is
decided to be the chosen drug. It is noteworthy that o
3
K
unequivocally delivers the same
judgement without any need to look into hesitation gures.
5. Conclusion
In order to establish Homeopathy as a true science at par with others, systematization
and standardization are prerequisites. A repertory of homeopathic symptoms-drugs is
one of the most important sources of the practice and research in this science. Hence
one reads now and then calls to develop new repertories based upon modern scientic
practices e.g. likelihood ratio [8] based repertory. Complementary to a new repertory is
a method of repertorization which may use the full benet of the new repertory. A two-
progend approach has been presented in this paper. On one hand a scheme to develop a
new repertory, establishing the associations and non-associations of a drug with a given
symptom has been presented. On the other hand a method to fully exploit the benets of
such a new repertory has been presented. For this purpose IFS theory has been used to
adapt Sanchezs approach for medical diagnosis ([9],[10]) to the problem of homeopathic
drug selection. Three dierent selection indices and two types of IFR compositions have
been discussed in the sequel. The method is then described in detail and applied to an
example case. Besides providing a methodology, we also propose an automated mechanism
to accumulate homeopathic knowledge in a systematic manner. This may help to realize
the such ideals as set in [8] and [12].
Acknowledgement: Author would like to gratefully acknowledge the valuable sug-
gestions of an anonymous referee and the editor, which helped to improve the presentation
of this paper, a lot.
References
[1] K. Atanassov, Intuitionistic fuzzy sets, Fuzzy Sets and Systems 20(1986) 87-96.
[2] K. Atanassov, G. Gargov, Intuitionistic fuzzy logic, C. R. Acad. Bulgare Sc. 43
(3)(1990) 9-12.
[3] K. Atanassov, G. Georgeiv, Intuitionistic fuzzy prolog, Fuzzy Sets and Systems
53(1993) 121-128.
[4] R. Biswas, Intutionistic fuzzy relations, Bull. Sous. Ens. Flous. Appl. (BUSEFAL)
70(1997) 22-29.
[5] H. Bustince, P. Burillo, Structures on intutionistic fuzzy relations, Fuzzy Sets and
Systems 78(1996) 293-303.
[6] W. L. Gau, D. J. Buehrer, Vague Sets, IEEE Trans. Sys. Man Cyber. 23(2)(1993)
610-614.
[7] J. T. Kent, Use of the repertory, in Repertory of the Homeopathic Materia Medica,
B. Jain Publishers India (1993).
Homeopathic Drug Selection Using Intuitionistic Fuzzy Sets 9
[8] A. Rutten, C. Stolper, R. Lugten and R. Barthels, Repertory and likelyhood ratio:
time for structural changes, Homeopathy 93(2004) 120-124.
[9] E. Sanchez, Resolution of composition of fuzzy relation equations, Information and
Control 30 (1976) 38-48.
[10] E. Sanchez, Solutions in composite fuzzy relation equation. Application to Medical
diagnosis in Brouwerian Logic, in: M.M. Gupta, G.N. Saridis, B.R. Gaines (Eds.),
Fuzzy Automata and Decision Process, Elsevier, North-Holland, 1977.
[11] M. H. Shu, C. H. Cheng, J. R. Chang, Using intuitionistic fuzzy sets for fault-tree
analysis on printed circuit board assembly, Microelectronics Reliability (Article in
press, doi:10.1016/j.microrel.2006.01.007).
[12] A. Steinsbekk and Data Colloection Group, Data collection in homeopathic practice:
A suggestion for an international standard, British Homeopathic Journal 89(2000).
[13] E. Szmidt, J. Kacprzyk, Intuitionistic fuzzy sets in group decision making, Notes on
IFS 2(1)(1996) 11-14.
[14] L. A. Zadeh, Fuzzy sets, Information and Control 8 (1965) 338 353.