Expert reflection in the process of diagnosis of diseases at the extraction of knowledge

Full text

Expert reflection in the process of diagnosis of
diseases
at the extraction of knowledge
Boris Kobrinskii
Laboratory Decision clinical support system
of Institute of Modern Information Technologies in Medicine of the Federal Research Center
“Computer Science and Control” RAS
Moscow, Russia
bak@isa.ru
Abstract—Reflection of the expert takes among in the process
of forming answers to the questions of the cognitive scientist. In
medical diagnostics this process including the analysis of disease
dynamics. This procedure can be presented in the form of a
formula that includes the fuzzy of pathological manifestations,
their relevant and the expert certainty factors. This factor
characterizes linguistic (verbal) knowledge and visual holistic
images in different time periods of the disease.
Keywords—formalized representation of the expert's reflection;
certainty factors; medical diagnostics; reflection of disease
dynamics; fuzzy of disease manifestations; constructs and
reflection.
I. INTRODUCTION
In the mathematical theory of reflexive structures Lefebvre
[1] considers the behavior of a person living in a +/- world,
where there are only two poles: positive and negative. A
person manifests an inner unconscious intention to a positive
or negative choice. The preferable choice is carried out under
the influence of the "image of yourself" and the "image of the
situation" that a person develops from his self-assessment and
evaluation of his concept of himself and the outside world.
The American writer Henry David Thoreau wrote about it
metaphorically [2]: "No matter how acute my experiences are,
I always feel that some part of me treats them critically; it's
not even part of me, but an observer who does not share my
experiences and only celebrates them".
Unlike the "image of oneself," the model of itself does not
have a reality status for the subject [3]. In the logic of
argumentation, as the shows [4], all the main results of V.A.
Lefebvre, however many interpretations of evaluations and
behavior expand due to the differentiation of indefinite and
positive action. Sets of arguments in the decision making are
divided into disjoint subsets, each of which reflects some
aspect of the value system. In the set of arguments, it is
possible to select, according to the modality principle, a lot of
knowledge.
The phenomenon of reflexion (in relation to medicine) is
the reflection (transformation) of the external world of the
disease in a particular person (the perception of pathological
manifestations) through the inner world of a person, through
the psychic sphere of the doctor and the expert doctor or
consultant.
II. REFLECTIVITY OF THE MEDICAL EXPERT
Reflection manifests itself not only in the attending
physician who makes the decision [5], but also with the
expert-physician doctor in the process of extracting knowledge
from him for creating an intelligent system for supporting the
adoption of diagnostic or therapeutic solutions. Let's consider
the problem of making diagnostic decisions in the future. Alter
ego not only analyzes and evaluates the first hypothesis, but
also forms alternative hypotheses, "proposes" proposals for
checking the diagnostic significance of information (signs)
and possible solutions caused by differential diagnosis with
similar diseases.
In transactional categories of description, individuality is
the cumulative potential of ego-states as a condition for self-
regulation [6
]. There is evidence that an innate
information processor is "built" into the human psyche, the
function of which is to automatically generate images of
oneself and others along with their subjective world. The
operation of this processor generates a specific spectrum of
human responses that are not consciously controlled and
proceed extremely fast (1 to 2 milliseconds). This kind of
reflection, unlike the traditional one, was called fast reflection
[7]. In certain cases, it is possible to observe a very fast
reaction of the expert to the questions of the cognitive scientist
with the presentation of complex diagnostic constructions.
At the same time, the tendency to generate a line of
behavior will be called the law of self-reflection [8]. This law,
which generates the line of the mental reasoning of the expert,
is defined in the application to medicine, on the one hand, by
an individual scheme of traditional diagnostic and therapeutic
This is an open access article under the CC BY-NC license (http://creativecommons.org/licenses/by-nc/4.0/).
Copyright © 2017, the Authors. Published by Atlantis Press.
Advances in Computer Science Research (ACSR), volume 72
(ITSMSSM 2017)
321
IV International Research Conference "Information Technologies in Science, Management, Social Sphere and Medicine"

processes, on the other hand, by the character and direction of
the questions of the cognitive scientist. These questions can
partially deform the habitual order of analyzing specific data
on the basis of knowledge that has been formed in the process
of a generally accepted diagnostic procedure and personal
experience (heuristic knowledge). In the presence of a
preliminary textual analysis of literary sources, experts take
into account subjectivity as available descriptions (severity of
symptoms, time of manifestation, etc.), and their assessments
from the standpoint of individual knowledge about this
pathology. With the group extraction of reflexive knowledge,
not only the presentations caused by questions of the cognitive
scientist, but also the statements of fellow experts are exposed.
In this case, the variables are additionally corrected in the
reflexive system of the main expert by the knowledge of other
experts about the phenomenon being studied (manifestation of
the disease).
III. IMAGES AND FUZZY IN MEDICAL PRACTICE
The model of a reflexive solution is not a utility
maximization model. It takes into account the polarity
(positive and negative quality) outside the context of utility.
But it can also take into account the degree of confidence in
the decision. The reliability of experts in the expression of
signs is manifested in assessing the similarity or difference
with the cognitive mental image, in the conviction or doubt in
their knowledge about the phenomenon under consideration,
which is due not only to the depth of knowledge, but also to
the reflexivity of the subject. In particular, this applies to
different figurative representations (situations as images based
on multi-component features, visual images). It should be
borne in mind that the image can be [9]: 1) based on the
memory (mental analogy); 2) the result of the simultaneous
effect (pseudo-semantic image); 3) associated with the
semantic concepts that generate the hypothesis of a diagnosis
when a pathognomonic symptom is found that "performs the
function of mapping the whole" is akin to a self-reflexive
system [10]; 4) the result of the "emergence" of the visual
image (pseudo-visual image of the expert).
Any image (semantic script or visual image) almost always
differs in fuzzy [11] or underdetermination (NON-factors by
A.S. Narinyani [12]), presented in the form of a linguistic
scale, taking into account both the temporal characteristics,
and his perception is always connected with reflection. In turn,
the physician's reflexion and scientific preferences
involuntarily lead to increased attention to certain information
(activation of certain signs and images) and to an artificial
lowering of the role of other data immersed in the depth of
consciousness [13]. These images, driven out of
consciousness, can subsequently "float up" and again
participate in decision-making. If the image seems
incomprehensible, if it irritates and does not finish up to
something acceptable from the point of view of equilibrium
and symmetry, then it is superseded. It exists somewhere in
the subconscious until the intuition happen, there is no insight
or insight, and then it becomes clear that this image explains
the imperfection of the other analyzed image. Such images
recessed to a certain time in the subconscious seem to be not
trivial (specific) holistic images with orphan (rare) diseases
that do not correspond to the usual notions and are not
verbalized without loss of integrity.
IV. CONSTRUCTS AND REFLECTION
According to G. Kelly [14], everything depends on how a
person realizes and interprets a particular phenomenon. It does
not matter what we are trying to understand, there are always
"constructive alternatives" open to our minds. According to
Kelly, people are researchers who, in relation to current events
and situations, construct hypotheses, and in the case of their
refutation - alternatives to these hypotheses. In medicine, they
are based on counter-signs or signs-negations. Constructs for
their implementation include a professional component and
personal perception, including reflection. A construct is a
special subjective tool designed by the person himself,
allowing him to perceive and understand the surrounding
reality, to evaluate and predict events. Constructs can be both
verbal (conscious constructs, which can be expressed by
words), and verbal (unconscious constructs, which are used if
a person does not have words to express it). To measure
personal constructs, a methodical principle of "repertory grids"
and a repertoire test of personal constructs was developed. It is
important to note the lability of repertory grids. This
phenomenon can be noted and the expert.
L.Hjelle and D.Ziegler [15] formulated the concept of the
individual's personality as an organized system of more or less
important personal constructs that a person uses to interpret
the world of experiences and to foresee future events, i.
dynamics of changes in time.
V. REFLECTIVE MODEL OF THE EXPERT
Such a model in the field of medical diagnostics can be
based on a scheme that includes a reflexive self-analysis of
differential diagnosis. This is a cognitive representation: in the
traditional form, a personal "image of the disease" on the basis
of precedent or analogues, intuitive notions about the features
of the clinical picture and the relationships between the signs,
the transformation of manifestations of the disease in time.
These representations are characterized by fuzzy
manifestations, represented in natural language in linguistic
scales with subsequent translation into numerical form [0, 1].
In the process of extracting knowledge from an expert (or
group of experts), the cognitive scientist asks him a series of
questions that are aimed at identifying and clarifying
knowledge about the disease(s) needed for diagnosis
(differential diagnosis). These questions may be pattern of
general, clarifying or sounding, and others, including control
questions. Integration of numerous sessions of work with an
expert allows you to form the field of knowledge of the
problem area. The results of the session, including the
reflexive component of experts, can be presented in a
formalized form.
It is possible to introduce a cortege (x) that reflects the
expert's view of a particular disease or a differential diagnostic
range of diseases during a session with a cognitive scientist,
including a fuzzy variable (y1,2, ..., n), characterizing a
number of manifestations of the disease, presented by the
expert in verbal form at the stage of knowledge extraction (for
Advances in Computer Science Research (ACSR), volume 72
322

fuzzy signs a specially formed linguistic scale is used). This
knowledge is formed by the expert in the process of reflexive
analysis of the answer to the questions of the cognitive
scientist and is accompanied by certainty factor [16] (cf) [-1,
+1] about each diagnostic feature. Certainty factor can refer to
either linguistic or verbalized knowledge in the form of
separate signs or symptom-complexes (which can be equated
to scenarios) or to holistic (integral) images of a visual type
for a particular disease in a certain age period. Depending on
the nature of the information analyzed, transitional options are
possible from almost absolute certainty, which characterizes
an extremely high degree of certainty, to the maximum
expressed doubt (an extremely low degree of certainty). In
addition, the expert's reflective system evaluates their
relevance (r), taking into account the dignity value: especially
important, important, relatively significant, little significant,
etc. Subsequently this fuzzy linguistic scale is translated by
the cognitive scientist into a numerical form in the range from
0.1 to 1).
Assuming that the variables (x, y, cf, r) are defined on the
set of real numbers and complemented by the discrete time
characteristic T1,2 ..., n, which reflects the changes in the disease
in the dynamics, we get a number of knowledge represented
by the expert as a function:


,
, ,
f
x y c r T
 
where x is a cortege reflecting the expert's presentation of a
particular disease or a differential diagnostic range of diseases,
y1,2, ..., n - fuzzy variable characterizing the signs and
holistic images of the disease,
cf - certainty factor of the expert in the degree of
expression of each feature,
r1,2, …, n - the relevance of each sign for the diagnosed
disease; T1,2, …, n - time (period) of disease, characterizing the
manifestation of signs.
Cortege (x) and the variables (y1,2, ..., n, cf, r1,2, ..., n)
depend on T due to the nature of the manifestation of signs
and images in time. When solving a diagnostic task only for a
specific point in time or a disease that does not manifest
changes in dynamics, the variable T is not included in the
above formula.
VI. THE CONCLUSION
The reflection takes place at all stages of perception /
processing of symptoms, symptom-complexes and images and
directs further search for a diagnostic solution along a certain
path. However, symptoms in medicine are often characterized
by a wide range of manifestations, which corresponds to the
concept of fuzzy. The same applies to the image, as well as to
the metaphor (for example, Münchausen syndrome in
psychiatry) used in medicine, which also differ in fuzzy, i.e.
variability of manifestations, and perception of the image or
metaphor is always connected with the subject's reflection.
This also takes place in the reflection of the logical-and-
argumentative constructions of the expert, traditionally
presented primarily in linguistic form. The analysis of
knowledges and datas in dynamics is determined by the age
periods of the diseases.
The process of the expert's cogitative activity as a whole,
including the expert's reflection, can be implemented in a
formalized form and presented on the basis of a formula that
includes a fuzzy variable, a confidence factor, relevance of
characteristics and a temporal characteristic of changes in the
course of diseases.
REFERENCES
[1] V. Lefebvre, “Choice without criteria of preference”, in The structure of
human reflexion: The reflexional psychology of Vladimir L efebvre, H.
Wheeler, Eds. New York: Peter Lang Publ, 1990, pp.93–112.
[2] H.D. Thoreau, “Wal den or life in the woods”, Moscow, Nauka, 1979,
456 pp. (in Rus).
[3] V.A. Lefebvre “The cosmic subject”. Moscow: Russian Academy of
Sciences, Institute of Psychology Press, 1997, 184 pp. (in Rus).
[4] Т.А. Taran, “The problems of conflict modeling and reflexive behavior:
the use of the logic of argumentation”, VI Intern. Conf. "Knowledge-
Dialogue-Solution" KDS-97: Collection of scientific papers, vol.1. –
Yalta, 1997, pp.47-56. (in Rus).
[5] B.A. Kobrinskii, “Reflection and fuzzy representations in medical
systems of artificial intelligence”, Reflexive control. Intern. sympos.
theses. Moscow: Russian Academy of Sciences, Institute of Psychology
Press, 1997, pp.85-86. (in Ru s).
[6] V.А. Petrovsky, “Individuality and self-regulation: a metamplicative
model”, in Reflective approach: from methodology to practice,
V.E.Lepsky, Eds. Moscow: Kogito-centre, 2009, pp.205-234. (in Rus).
[7] V.A. Lefebvre, “The Fundamental Structures of Human Reflection”, J.
Soc. Biol. Struct., London, New York, vol.10, iss.2. pp.129-175, April
1987. (references)
[8] V.A Lefebvre, “The Law of Self-Reflexion: A possible unified
explanation for three different psychological phenomena”, Refl.
Processes and Control, vol.1, No.2. pp.91-99, 2002. (references)
[9] B.A. Kobrinskii, “Reflection of image thinking in artificial intelligence
systems”, VI Intern. Conf. "Knowledge-Dialogue-Solution" KDS-97:
Collection of scientific papers, vol.1. – Yalta, 1997, pp.29-36. (in Rus).
[10] V.A Lefebvre, “About self-organizing and self-reflexive systems and
their research”, Applied ergonomics: Special issue: Reflexive processes.
Moscow, no.1, pp.6-9, 1994 (references) (in Rus).
[11] L.A. Zadeh “Fuzzy sets”, I nform. and Control., vol.8, iss.3, pp.338 -353,
June 1965. (references)
[12] A.S. Narinyani “Underdetermination in the system of representation and
processing of knowledge”, Izv. AN SSSR: Techn. Cyber., Moscow,
no.5, pp.3-28, 1986. (references) (in Rus).
[13] D.A. Pospelov “Metaphor, image and symbol in the knowledge of the
world”, News of artif. Intel., Moscow, no.1, pp. 91-114, 1998.
(references) (in Rus).
[14] G. Kelly “Personal Theory (Personal Construct Theory)“, St. Petersburg,
Rech, 2000, 249 pp.
[15] L. Hjelle, D. Ziegler “Per sonal T heory”, 3rd intern. ed., St. Petersburg,
Piter, 2003, 608 pp.
[16] D. Heckerman, E.H. Shortliffe “From certainty factors to belief
networks”, Artif. Intell. Med., vol.4, iss.1, pp.35-52, February 1992.
(references)
Advances in Computer Science Research (ACSR), volume 72
323