DEFINITIONS OVERVIEW OF FUZZY ONTOLOGY
Hai Bang Truong, and Hoang Thanh Duong University of Information Technology, VNU HCM
ABSTRACT
The conceptual formalism supported by ontology is not sufficient for handling vague information that is commonly found in many applications. Fuzzy ontology is an extension of the domain ontology for solving the uncertainty problems. Current fuzzy ontology models do not focus on essential semantic relationships between fuzzy concepts, which lead difficulty in ontol
DEFINITIONS OVERVIEW OF FUZZY ONTOLOGY
Hai Bang Truong, and Hoang Thanh Duong University of Information Technology, VNU HCM
ABSTRACT
The conceptual formalism supported by ontology is not sufficient for handling vague information that is commonly found in many applications. Fuzzy ontology is an extension of the domain ontology for solving the uncertainty problems. Current fuzzy ontology models do not focus on essential semantic relationships between fuzzy concepts, which lead difficulty in ontol
Copyright:
Attribution Non-Commercial (BY-NC)
Available Formats
Download as DOCX, PDF, TXT or read online from Scribd
DEFINITIONS OVERVIEW OF FUZZY ONTOLOGY
Hai Bang Truong, and Hoang Thanh Duong University of Information Technology, VNU HCM
ABSTRACT
The conceptual formalism supported by ontology is not sufficient for handling vague information that is commonly found in many applications. Fuzzy ontology is an extension of the domain ontology for solving the uncertainty problems. Current fuzzy ontology models do not focus on essential semantic relationships between fuzzy concepts, which lead difficulty in ontol
Copyright:
Attribution Non-Commercial (BY-NC)
Available Formats
Download as DOCX, PDF, TXT or read online from Scribd
LOGICS The conceptual formalism supported by ontology 2.1. Definition of Ontology is not sufficient for handling vague information that is commonly found in many applications. Fuzzy Definition 1[12] ontology is an extension of the domain ontology for Most often ontology is defined as a structure solving the uncertainty problems. Current fuzzy consisting of the following elements (Gruber 1993; ontology models do not focus on essential semantic Fensel 2001): relationships between fuzzy concepts, which lead C – a set of concepts (classes); difficulty in ontology integrating. To represent I – set of instances of concepts; formally the fuzzy knowledge, this paper presents a R – set of binary relations defined on C; series Definitions of fuzzy ontology and issues related. Z – set of axioms which are logic formulas representing integrity constraints in the sets of Keywords: Ontology, Fuzzy ontology, Fuzzy instances and concepts. system.
1. INTRODUCTION Definition 2[1]
In the last few years, the application of ontologies An ontology is a 4-tuple O = (C, P, R, A), where: as formalisms for knowledge representation in many C is a set of concepts defined for the domain. A different application domains has grown significantly. concept is often considered as a class in ontology. An ontology is defined as an explicit and formal P is a set of concept properties. A property p ∈ P is specification of a shared conceptualization, which defined as an instance of a ternary relation of the means that ontologies represent the concepts and the form p(c, v, f), where c ∈ C is an ontology concept, relationships in a domain promoting interrelation with v is a property value associated with c and f other models and automatic processing. Ontologies defines restriction facets on v. Some of the have a lot of advantages, such as making possible to restriction facets are – type (ft), cardinality (fc), and add semantics to data, making knowledge range (fr). The type facet ft may be any one from maintenance, information integration as well as the the standard data types supported by ontology reuse of components easier. editors, i.e. ft ∈ {boolean, integer, float, string, symbol, instance, class, …}. The cardinality facet Ontologies are formal descriptions of some fc defines the upper and lower limits on the number domain’s vocabulary. They allow adding semantic of values for the property. The range facet fr descriptions to data, which makes knowledge specifies a range of values that can be assigned to maintenance, information integration and the reuse of the property. components easier. Significant use as formalisms for Knowledge Representation, they have been R = {r | r C × C × Rt} is a set of binary semantic successfully used in expert and multi-agent systems, in relations defined between concepts in C. Rt = the Semantic Web… Their main theoretical formalism {one-to-one, one-to-many, many-to-many} is the is Description Logics (DL) [5] [9]. Current standard set of relation type. language of ontology is OWL, Web Ontology A is a set of axioms. An axiom is a real fact or Language, which is based on a DL. Several extensions reasoning rule. to OWL have been proposed. Among them, OWL 2 2.2. Description Logics seems to become its successor [11]. Description Logics (DLs) are a family of logics for representing structured knowledge. From a historical point of view, DLs were born to provide a semantic for semantic works and frames, two classical models of structured knowledge representation. In order to overcome their lack of semantics, it was observed that they could be given a semantic by using First Order Logic (FOL). DLs are a family of logics for structured knowledge representation, born to give a semantic to frames and semantic networks. Usually, they are fragments of First Order Logic with binary relations.
3. FUZZY ONTOLOGY
3.1. Fuzzy Ontology
Various Definitions of Fuzzy Ontology have been introduced, although we think that neither of them is general enough. A list of the most relevant definitions is included here, where the notation has been Fig. 1: Structure of fuzzy ontology homogenized and, in some cases, summarized. It is worth to note, however, that some of the authors do not claim to be proposing a universal definition of the term. Definition 2[1]: i. Fuzzy domain ontology: Definition 1[2] A fuzzy domain ontology is a 4-tuple OF(C, PF, i. Domain Ontology: RF, AF) where: A domain ontology defines a set of C is a set of concepts. Every concept here has representational terms that we call concepts. Inter- some properties whose value is fuzzy concept or relationships among these concepts describe a target fuzzy set. world. There are four layers, including domain layer, PF is a set of properties. A property defined as a category layer, event layer, and class layer, defined in 5-tuple of the form PF(c, vF, qF, f, U), where c ∈ C the domain ontology. The domain layer represents the is an ontology concept, vF represents property domain name of an ontology and consists of various values, qF models linguistic qualifiers, which can categories defined by domain experts. Each category is control or alter the strength of a property value vF, composed of the event set, which is derived from the f is the restriction facets on vF, and U is the news corpus by domain experts. Every event universe of discourse. Both vF and qF are the fuzzy comprises several concepts of class layer. In class concepts at U, but qF changes the fuzzy degree of layer, each concept contains a concept name, an vF. attribute set, and an operation set for an application RF is a set of inter-concept relations between domain. There are three kinds of relationship, concepts. Like fuzzy concept properties, rF ∈ RF is including generalization, aggregation, and defined as a 5-tuple of the form rF(c1, c2, t, sF, U) association, in the domain ontology. The relationship where c1, c2 ∈ C are ontology concepts, t between a domain and its corresponding category is represents relation type, U is the universe of generalization that represents “is-kind-of” relationship. discourse, and sF models relation strengths and is The relationship between each category and its fuzzy concept at U, which can represent the corresponding events is aggregation. The aggregation strength of association between concept-pairs c1, denotes “is-part-of” relationship. The association c2. represents a semantic relationship between concepts in AF is a set of fuzzy rules. In a fuzzy system the set class layer. of fuzzy rules is used as knowledge base. The fuzzy domain ontology is used to model domain ii. Fuzzy Ontology: expert knowledge. But, due to the lack of A fuzzy ontology is an extended domain ontology relationships between fuzzy concepts that can be with fuzzy concepts and fuzzy relationships. Fig. 2 the value of properties, it is difficult to integrate shows the structure of the fuzzy ontology: diverse ontology systems. For example, in an ontology the set of property “price” value is fuzzy concept that can be the value of ca , e.g. {cheap, appropriate, expensive, …}, and in other “very cheap”. ontology the same set is {high, low, middle, …}. O is the set of fuzzy operators at U, which is To map these ontologies, it is necessary to define isomorphic to Q. the semantic relationship between fuzzy concepts, e.g. “cheap” and “expensive” have the relation of v. Basic fuzzy ontology disjointness, and “low” and “high” have the same A basic fuzzy ontology is a 4-tuple OF = (ca, CF, relation of disjointness etc. F, U), where: ca, CF, F, U has same interpretations as defined in definition (iii), which satisfy the following ii. Fuzzy linguistic variable conditions: Fuzzy linguistic variable is the variable whose CF = {c1, c2,… , cn} is a limited set. value is term or concept in natural language. A fuzzy Only one relation of set, the relation of linguistic variable is a 4-tuple (X, T, M, U) where: disjointness, exists in CF, and CF is complete at U. X is the name of fuzzy linguistic variable, e.g. In the other words, CF is a fuzzy partition of U. “price” or “speed” etc. CF has an ordered relation ≤, and CF, ≤ is a T is the set of terms which is the value of fuzzy complete ordered set, i.e. all concepts in CF linguistic variable, e.g. T = {cheap, appropriate, constitute a chain n c1 ≤ c2 ≤…≤ cn. expensive,…} or T = {fast, middle, slow,…}. F is optional element of ontology. M is the mapping rules which map every term of T to fuzzy set at U. An example of basic fuzzy ontology is OF = (ca U is the universe of discourse. Introducing = price of fruit, CF = {very cheap, cheap, appropriate, semantic relationships between concepts we expensive, very expensive}, U = [0,100]). obtain the ontology model.
iii. Fuzzy linguistic variable ontology
A fuzzy linguistic variable ontology is a 5-tuple OF = (ca, CF, R, F, U) where: ca is a concept on the abstract level, e.g. “price”, “speed” etc. CF is the set of fuzzy concepts which describes all values of ca. R = {r | r ⊆ CF × CF} is a set of binary relations between concepts in CF. A kind of relation is set relation RS = {inclusion, intersection, disjointness, complement}, and the other relations are the order relation and equivalence relation RO = {≤, ≥, =}. CF and an order relation r compose the ordered structure CF, r. Fig. 2: Three-layered ontology structure F is the set of membership functions at U, which is isomorphic to CF. U is the universe of discourse. Definition 3 [5, 3] A Fuzzy Ontology is defined as the 5-tuple OF = iv. Extended fuzzy ontology {I, C, R, F, A} where: An extended fuzzy ontology is a 7-tuple OF = (ca, I is the set of individuals, also called instances of CF, R, F, Q, O, U), where: the concepts. ca, CF, R, F, U have same interpretations as C is the set of concepts. Each concept c ∈ C is a defined in (iii). fuzzy set on the domain of instances C: I → [0, 1]. Q is the set of the linguistic qualifiers, e.g. Q = The set of entities of the fuzzy ontology will be {very, little, close to, …}. An qualifier q ∈ Q and a indicated by E, i.e., E = C ∪ I. fuzzy concept F, CF ∈ C compose a composition R is the set of relations. Each r ∈ R is an n-ary fuzzy set is defined over a domain that overlays fuzzy relation on the domain of entities, R: En → part of the universe of discourse. [0, 1]. A special role is held by the taxonomic relation T: E2 → [0, 1] which identifies the fuzzy Definition 6 [15] subsumption relation among the entities. The fuzzy ontology is based around the concept F is the set of the fuzzy relations on the set of that each index term or object is related to every other entities E and a specific domain contained in D = term (or object) in the ontology, with a degree of {integer, string, ...}. In detail, they are n-ary membership assigned to that relationship based on functions such that each element f ∈ F is a relation fuzzy logic as introduced by (Zadeh, 1965). The fuzzy F: E(n−1) × P → [0, 1] where P ∈ D. membership value µ is used for the relationship A is the set of axioms expressed in a proper logical between the term or object in question where 0< µ <1 language, i.e., predicates that constrain the and µ corresponds to a fuzzy membership relation such meaning of concepts, individuals, relationships as “strongly”, “partially”, “somewhat”, “slightly” and functions. where for each term: i=n Definition 4 [6] ∑ μ i=1 A fuzzy ontology is an ontology extended with i=1
fuzzy values which are assigned through the two
Where n is the number of relations a particular functions g: (Concepts Instances) × (Properties object has, where n = (N - 1), with N representing the Pro_val) → [0; 1] and h: Concepts Instances → [0; total number of objects in the ontology. That is, each 1]. term used in the system has the total membership value Definition 5 [8] of its relations as a value of 1 summed over each dependant relation. This rule is not commutative, for A Fuzzy Ontology OF, is a quadruple of the form the relationship between two objects, A, B: OF = (C, ΡF, F, M), where: C is a set of concepts defined for the domain. A related ¿ B … μ AB → ΡF is a set of fuzzy concept properties. A property p P is defined as a quadruple of the form pf(c, vf, B related ¿ A … μBA → qf, f), where c C is an ontology concept, ‘vf’ µAB > µBA or µAB < µBA or µAB = µBA are all possible. For represents fuzzy attribute values and could be simplicity, it is assumed there is only one type of either fuzzy numbers or fuzzy quantifiers, ‘qf’ relation between the objects A and B. For further models linguistic qualifiers and are hedges, which simplicity, for each membership value for a relation can control or alter the strength of an attribute originating from A, the membership value is written as value and f is the restriction facet on vf. µB rather than µAB as the A can be assumed. The overall F is a set of fuzzy inter-concept relations value of 1 is used in order to prevent objects with large between concepts. Like fuzzy concept properties, numbers of relations having a larger effect in a search F is defined as a quadruple of the form F (ci, cj, process than those with relatively small numbers of t, qf), where ci, cj C are ontology concepts, ‘t’ relations. represents relation type, and ‘qf’ models relation strengths and are linguistic variables, which can Imagine that such a constraint does not apply. represent the strength of association between Consider a query with two terms A and B. If A has only 2 relations, with µ1 = 0.7 and µ2 = 0.6 then ∑ = 1.3. concept-pairs ci, cj. However, if B has 6 relations say, µ1 = 0.7 and µ2 = The choice of fuzzy numbers or fuzzy quantifiers 0.5, µ3 = 0.7 and µ4 = 0.6, µ5 =0.7 and µ6 = 0.6 then ∑ for values is dictated by the nature of the = 3.8. This doesn’t matter if A and B are the only underlying attribute and also its restriction facets. elements in the query and both are needed, but if there The complete range of values over which an is a list of terms, so that all of them is unlikely to attribute can take values defines the universe of retrieve wanted documents then there will be a discourse M. The universe of discourse is tendency to select B as part of the final query rather decomposed into a collection of fuzzy sets. Each than A and hence a bias will occur, and there will be a tendency to expand via related terms of B. Obviously being used for query construction. In addition, this argument can be reversed, so that relatives of A unwanted cyclic relationships may occur. Unwanted could be seen as having too high a membership value, relationships can be assigned zero membership values. but it is likely that in a well designed ontology, the In practice, unused terms or objects may also be relatives of very connected terms will have sufficiently ignored in the ontology, as there is no need to retain distinct meanings to be related but unwanted. them to act as a bridge between otherwise useful Practically, in addition, if a threshold value of µ is terms. used, large numbers of relatives will not qualify and so huge query expansion will not occur. 4. DISCUSSION These definitions try to formalize the notion of This does mean that there is not a consistent fuzzy ontology by means of an enumeration of the mapping between the strength of a relationship in elements of the ontology which are extended in order words and the membership value, but for each object, to support vague knowledge. Despite the merit of these the strongest relationship will continue to have the definitions, this approach has many problems highest membership value. discussion. Typically, different domains will need to Computer represent vagueness and imprecision at different Tree Food company levels. Furthermore, future languages will offer new µ=0.2 is a is a possibilities to be extended which are unknown µ=0.3 µ=0.09 nowadays, but these definitions do not allow other product of µ=0.01 fuzzy elements than the explicitly mentioned. As a Apple manufactures consequence, the scalability and reusability of the Granny is a iPod definition are compromised. For example, a fuzzy role Smith µ=0.2 hierarchy falls out of these definitions since taxonomic is a relations between roles are not mentioned. Sµ=1 µ=0.2 Pippin
Most of the approaches are even more restrictive
Fig. 3: Example Fuzzy Ontology Scheme and are dependent application dependent, since they propose the minimal extensions which are sufficient to An example, Figure 3, may take this clearer, each cover a particular application. For example, do not µ represents the membership value of the relationship fuzzy concepts, whereas does not fuzzy relations. from the apple to tree, fruit and computer company. Moreover, some definitions are tied to a particular Any relationships not shown, e.g. between IPOD and formalism. For example is attached to the fuzzy pippin, are assumed to have a µ = 0, relationships language FUZZY OWL based on a fuzzy extension of directed to “Apple” do not have µ values shown for the DL SHOIN. But even if it became a standard clarity. language for fuzzy ontologies representation, there are still some interesting features which cannot be The apple can be a product of a tree (to distinguish represented in it, such as fuzzy nominals or fuzzy an apple tree from another kind of tree); it can be a concepts or role hierarchies. We understand fuzzy fruit (if the query relates to distinguishing between ontologies in a more general sense. In our opinion, a fruit, for example apple pie, cherry pie) or a computer fuzzy ontology is simply an ontology which uses fuzzy company. The fuzzy ontology applies membership logic to provide a natural representation of imprecise values to each of these possibilities; depending on how and vague knowledge, and eases reasoning over it. likely it is that a particular relation is required. These different relations will have different membership Typically, concepts, (abstract and concrete) values depending on the context of the query, and relations and axioms can be fuzzified. But fuzziness particularly the user’s view of the world. The can also be in fuzzy concrete domains or in external capitalization of the word Apple/apple may also be formalisms such as a fuzzy rule layer. Our informal used to assist this differentiation, but most web-based definition is general enough to allow the ontologists to searching tools seem to ignore this information. decide in which levels they want to introduce the fuzziness, makes possible to have a crisp ontology In conventional ontologies, particular objects may with fuzziness being dealt with using a external occur in multiple locations, leading to ambiguity when formalism, and, what seems more important to us, will ontology framework is common, just the membership not be compromised by the apparition of future values are different. ontology languages. Finally, this approach holds out the possibility It is also worth to recall that since fuzzy logic is a that the representation of a potentially very large generalization of classical logic, classical ontologies ontology can be compressed. If whole areas are not are special case of fuzzy ontologies, so this formalism required, the relations to the core can be set to zero. is by nature backwards compatible with current crisp Unwanted intermediate levels can also be removed, ontologies. with lower-level terms only communicating directly with higher levels. This aspect removes the need to Finally, we note in passing that, although in the create artificial groupings to avoid orphaned terms. At Artificial Intelligence literature the term approximate the limit a fuzzy ontology, with all membership values reasoning is directly related to fuzzy logic, in the set to 0 or 1, will have each term or object having one ontology literature this term has nothing to do with relation only. If each term only has one relation then a fuzziness. B-tree structure is possible, with each term only relating to its parent, however this arrangement is more 5. ADVANTAGES OF THE FUZZY properly called a hierarchy. Table 1 summaries some ONTOLOGY [15] of the differences between crisp and fuzzy ontology. Many issues arise from the use of multiple ontologies including the difficulties associated with Problems Fuzzy Ontology Crisp Ontology communicating between ontologies and the need for Compared maintenance of large numbers of ontologies. The Multiple- Does not occur Issue for Located terms disambiguation fuzzy ontology as described is partly suggested in Query Depends on Depends on order to allow a common framework, or base ontology, expansion membership location only with different membership values associated with value different users and groups. Customization Simple, based on Requires new modification ontology and/or Another advantage of this approach is membership ontology sharing completeness. Rather than impose an arbitrary values standard of the importance of a particular location in Storage Depends on the Depends on the ontology, which is required in a crisp ontology to required number of term number of terms avoid too many examples of a term appearing in the in the ontology in the ontology and the ontology, the term or object can be located in all membership relevant locations. values of the relations, can be Most importantly, for searching processes, the use smaller or larger of a fuzzy ontology for the mapping of search terms than crisp allows the relative weight of each term in the required Knowledge Related to use Related to output to be calculated. By allowing these weights to representation structure. be calculated accurately, it removes the bias associated Table 1: Comparison between Fuzzy Ontology and with multiply located terms being used for searching. Crisp Ontology If a term is located in multiple locations in a crisp ontology, and is used for query expansion purposes – say by including offspring, then the danger is that the large number of relatively irrelevant expansion terms 6. APPLICATIONS outweigh those which are useful. One of the most important applications is Semantic Web [4] and, more generally, the Internet, In particular, the use of a fuzzy ontology approach Electronic Commerce [9], Knowledge Management allows the convenient representation of the [1]. The World Wide Web Consortium (W3C) recently relationship in a domain according to a particular view, set up an incubator group on Uncertainty Reasoning without sacrificing commonality with other views; the for the Web, where “uncertain is intended to encompass a variety of aspects of imperfect knowledge, including incompleteness, [5] Silvia Calegari and Davide Ciucci, Fuzzy inconclusiveness, vagueness, ambiguity, and others” . Ontology, Fuzzy Description Logics and Fuzzy-OWL, Dipartimento di Informatica Sistemistica e Another important application is information Comunicazione Universit`a di Milano – Bicocca retrieval [10]. More specifically, medical information (Italia). retrieval [12], data mining [11]… have been the [6] Silvia Calegari and Davide Ciucci, Interrating subject of considerable research. Fuzzy logi In Ontologyes, Dipartimento di Informatica, Sistemistica e Comunicazione, Universit`a degli Studi di Milano Bicocca, via Bicocca degli Arcimboldi, Milano, Italy. [7] Hua-Mao Gu1, Xun Wang, Yun Ling, and Jin-Qin Shi, Building a Fuzzy Ontology of Edutainment Using OWL, College of Information, Zhejiang Gongshang University, Hangzhou, China. [8] Muhammad Abulaish, Member, IEEE & IEEE-CS, Department of Mathematics, Jamia Millia Islamia (A Central Univ.) Jamia Nagar, New Delhi – 25, India. Lipika Dey, Member, IEEE & IEEE-CS, A Fuzzy Ontology Generation Framework for Handling Uncertainties and Nonuniformity in Domain Fig. 4: Fuzzy Ontology-Classes Tab: Asserted Class Knowledge Description, Department of Mathematics, Hierarchy Indian Institute of Technology Hauz Khas, New Among the other examples that appear in the Delhi , India, literature we may cite multilingual ontologies, [9], Jun Zhai, Yiduo Liang, Yi Yu and Jiatao Jiang, ontology mapping, ontology integration, ontology Semantic Information Retrieval Based on Fuzzy dynamics, Chinese news summarization [2], and Ontology for Electronic Commerce. School of educational computer games [16]. Regarding fuzzy Management, Dalian Maritime University, Dalian, P. ontology editors, there is KAON editor, which presents R. China. an extension of the ontology[3]. There are also Fuzzy [10] Silvia Calegari and Elie Sanchez, Dipartimento Di extensions of Semantic Web languages such as RDF Informatica, Sistemistica e Comunicazione Universit`a [13], OWL, OWL 2 [5,7,14]. di Milano – Bicocca V.le Sarca, Milano (Italia) LIF, Biomathematiques et Informatique Medicale Faculte 7. REFERENCES de Medecine (Universite Aix-Marseille II) Bd Jean Moulin, Marseille Cedex, A Fuzzy Ontology-Approach [1] Jun Zhai, Lixin Shen, Zhou Zhou, Yan Liang, to improve Semantic Information Retrieval Fuzzy Ontology Model for Knowledge Management, [11] Eduardo L. G. Escovar, Cristiane A. Yaguinuma, School of Economics and Management, Dalian Mauro Biajiz, Using Fuzzy Ontologies to Extend Maritime University, Dalian P. R. China. Semantically Similar Data Mining, Department of [2] C.S. Lee, Z.W. Jian, L.K. Huang, A fuzzy ontology Computer Science – Federal University of S˜ao Carlos and its application to news summarization, IEEE (UFSCar). Transactions on Systems, Man, and Cybernetics, Part [12] DavidParry, Fuzzy Ontology for Medical B: Cybernetics 35 (5) (2005) 859–880. Document Retrieval, Proceedings of the 11th [3] Silvia Calegari and Davide Ciucci, Fuzzy Ontology International ConferenceonI nformation Pro-Cessing and Fuzzy-OWL in the KAON Project, Dipartimento and Managment of Uncertainty in Knowledge-Based Di Informatica, Sistemistica e Comunicazione Systems (IPMU2006).Paris, France2006.Pages2276– Universith di Milano - Bicocca. 2283. [4] Jun Zhai, Yan Chen, Qinglian Wang, Miao Lv, [13] Giorgos Stoilos, Giorgos Stamou, Vassilis Fuzzy Ontology Models Using Intuitionistic Fuzzy Set Tzouvaras, Jeff Z. Pan, and Ian Horrocks. “Fuzzy for Knowledge Sharing on the Semantic Web, School OWL: Uncertainty and the Semantic Web”. Fuzzy of Economics and Management, Dalian Maritime OWL: Uncertainty and the Semantic Web, In: University, Dalian, P.R. China. Proceedings of the 1st International Workshop on OWL: Experience and Directions (OWLED 2005), Volume 188, Galway, Ireland: CEUR Workshop Proceedings, 2005. [14] Fernando Bobillo and Umberto Straccia, Dpt.of Computer Science & Systems Engineering, An OWL Ontology for FuzzyOWL2. University of Zaragoza, Spain Istituto di Scienzae Tecnologie dell’ Informazione (ISTI-CNR), Pisa ,Italy. [15] David Tudor Parry, Fuzzy Ontology and Intelligent System for Discovery of Useful medical Information, a thesis submitted to Aucklend University of Technology in Fulfillment the degree of Doctor Philosophy, 2005. [16] website: http://www.w3.org/TR/owl2-profiles/ [17] Hanêne GHORBEL, Afef BAHRI, Rafik BOUAZIZ, Fuzzy Protégé for Fuzzy Ontology Models, MIRACL Laboratory, Faculty of economics and management of Sfax – Tunisia.