Lakhasly

Possibilistic Medical
Knowledge Representation
Model
MOHAMMAD ALSUN
LAURENT LECORNU
BASEL SOLAIMAN
Medical Decision Support Systems involve two main issues: medical knowledge representation and reasoning mechanisms adapted
to the considered representation model. This paper proposes an approach to construct a new medical knowledge representation model,
based on the use of possibility theory. The major interest of using
the possibility theory comes from its capacity to represent different types of information (quantitative, qualitative, binary, etc.), as
well as different forms of information imperfections such as uncertainty, imprecision, ambiguity and incompleteness. Starting from
the description, realized by an expert of the medical knowledge,
describing the relationship between symptoms and diagnoses, the
proposed approach consists on building a possibilistic model including the Medical Knowledge Base. Moreover, the proposed approach
integrates several possibilistic reasoning mechanisms based on the
considered knowledge. The validation of the proposed approach is
then conducted using an Endoscopic Knowledge Base. The proposed
representation, reasoning model and the obtained validation results
show a real interest in order to realize various goals of Medical Decision Support Systems such as classification, similarity estimation,
etc.
Manuscript received May 30, 2011; released for publication September 6, 2011.
Refereeing of this contribution was handled by Jean Dezert.
Author’s address: Institut Tel´ ecom, T ´ el´ ecom Bretagne, ITI Dpt, ´
Technopole Brest-Iroise, CS 83818, 29238 Brest Cedex 3, France. ˆ
1557-6418/12/$17.00 °c 2012 JAIF

1. INTRODUCTION
   Physician is the direct responsible for health and
   life of his patients. Therefore, diagnosis delivering is
   an extremely critical, although difficult task. Furthermore, diagnosis delivering is an error-prone task [3].
   Medical Decision Support Systems such as Knowledge
   Based Systems, Case Based Reasoning Systems, Machine Learning Systems and Medical Data Mining Systems, have been constructed in order to reduce diagnosis
   error risks, as well as to help physicians making high
   quality and reliable medical decisions [4]. These systems involve two main issues: the medical knowledge
   representation and adapted reasoning mechanisms. The
   medical knowledge, in general terms, has to be considered from two points of view: Expert Knowledge
   related to the physician’s description of different relationships between symptoms and diagnoses, symptoms
   and symptoms, and diagnoses and diagnoses. Patient
   Information is collected from each patient (patient data
   collecting and structuring). The first is crucial in order
   to establish Medical Knowledge Base, while the second
   leads to establish the patient database (i.e., Medical Case
   Base). Experts use their own experience of the medical
   cases as well as references knowledge sources to define
   the structure of the medical knowledge base.
   Medical knowledge often suffers from different
   forms of information imperfections (i.e., uncertainty,
   imprecision, ambiguity, etc.). In addition to the different types of information imperfections, the information
   can be quantitative (numerical or binary) or qualitative
   (nominal and ordinal) [17, 29]. Thus, the heterogeneity
   and imperfection of medical knowledge must be taken
   into consideration while the construction of a Medical
   Decision Support System. In other words, Medical Decision Support System has to be able to deal with heterogeneous and imperfect knowledge.
   In [27] R. Seising et al. defined the Medical knowledge as follows:
   “The certain information about relationships that exist between symptoms and symptoms, symptoms and diagnoses, diagnoses and diagnoses and more complex
   relationships of combinations of symptoms and diagnoses to a symptom or diagnosis are formalizations of
   what is called medical knowledge.”
   The term “symptom” is used for any information about
   the patient’s state health (anamnesis, signs, laboratory
   test results, etc.).
   According to the previous definition, the term “medical knowledge” will be considered in this study to
   represent the relationship between symptoms and diagnoses, (Symptom)—(Diagnosis). This relation is generally expressed in a probabilistic way based on the use
   of a linguistic term, referring to the expert’s assessment
   of the occurrence of a given symptom related to a given
   diagnosis.
   In order to be exploited in Medical Decision Support
   Systems, this Medical Knowledge has to be modeled
   (translated into a model understandable by the system)