Presentation held by K.U. Danyaro, J. Jaafar, and M.S. Liew at the Agricultural Ontology Service (AOS) Workshop 2012 in Kutching, Sarawak, Malaysia from September 3 - 4, 2012
1. Universiti Teknologi PETRONAS
Department of Computer & Information Sciences
Seri Iskandar, 31750 Tronoh, Perak, Malaysia
Fuzzy OWL-2 Annotation for MetOcean
Ontology
International Symposium on Agricultural Ontology
Service 2012 (AOS2012)
3 to 4 September 2012
Authors: K. U. Danyaro, J. Jaafar, and M. S. Liew
3. Motivations
Description logic (DL) is family of formal knowledge representation
language that has expressive power in reasoning concepts [1].
Provides logical formalism for ontologies and the Semantic Web
Needs fuzzy representation in order to meet the real world ontology
system.
Fuzzy DL are presented by extending classic DL to support the imprecise information
processing in ontology systems.
OWL needs to be used for representing the knowledge of a specific concept.
Meteorological and oceanographic (MetOcean) environment is also an
appropriate place to represent the knowledge based on fuzzy ontologies.
5. Introduction
Description Logic (DL)
A fragment of First-Order Logic (FOL).
Tarski-style declarative Semantics that enable capturing the standard
knowledge representation[2].
Is standardized by W3C standard for OWL Semantic Web (currently OWL
2) as the KR formalism.
Logic
Ontology
Computation
6. Introduction
An ontology is a formal explicit specification of a shared conceptualization
of a domain of interest [3, 4].
Description logic is usually employed to represent the knowledge and logic
of an ontology.
OWL helps in making connections between human and machine through
the logic concepts
Latest standardized OWL is OWL-2 which has a good feature for interaction between
machine and human i.e. Annotation.
“OWL is a computational logic-based language such that knowledge expressed in OWL can
be reasoned with by computer programs either to verify the consistency of that knowledge
or to make implicit knowledge explicit”[5].
OWL 2 has three important properties: object property, datatype property, and annotation
property.
8. MetOcean Dataset
MetOcean in situ contains large amount of data supplied by Cerigali
PETRONAS Sdn Bhd which is an Asia branch of MetOcean Company.
The 104.2e longitude and 5.45n latitude of Kota Kinabalu, a Malaysian
region of MetOcean has been used.
The time series data for the 2005 year, ranges from 1st January 2005 00:00
to 31st December 2005 23:00 was extracted using OSMOSIS software.
The typical hindcast data have been resulted in an array format.
The data spanned based on: YYYYMM, DDHH, WD, WS, ETOT, TP, VMD,
ETOT1, TP1, VMD1, ETOT2, TP2, VMD2 and HSIG.
Set all the datasets in form of OWL-2 relationships and particularized on
fuzzy variables (fuzzy elements for uncertainty and imprecision of the data).
10. Fuzzy OWL 2 Annotation
Fuzzy Interpretation
The convention of a statement in fuzzy logic is either true or false, 0 or
1.
⇒ the degree of truth of a statement ϕ is in the interpretation I.
⇒ fuzzy statement can be within f ∈ [0, 1], ϕ ≥ f or ϕ ≤ f , ϕ is a
statement
Definition: Let x be an element of ∆I (Interpretation domain) and .I be
the fuzzy interpretation function then the fuzzy interpretation I is a pair,
I = (∆I, .I) such that
– for every individual x mapped onto an element xI of ∆I,
– for every concept C mapped onto CI : ∆I → [0, 1],
– for every role R mapped onto a function RI: ∆I × ∆I → [0, 1].
Fuzzy interpretation I maps each statement into [0, 1], i.e. ∆I ⟶ [0, 1]
11. Fuzzy OWL 2 Annotation
Example
Wind Direction (deg)
Implies that the interpretation of very high can be determined by f I: ∆ID → [0, 1]. Where D is
a datatype property with <∆ID , 𝛗D>; ∆ID is the interpretation domain and the set of fuzzy
predicate.
Entailment equation as [f1 , f2] ⊆ Τ
Trapezoidal (f1 , f2, a , b, c, d), triangular (f1 , f2, a , b, c), right (f1 , f2, a , b) and left (f1 , f2, a , b).
f1 = 0, f2 = 1 is the modified datatype.
12. Fuzzy OWL 2 Annotation
Example
Datatype HighWindDirection: [0, 250] ⟶ [0, 1] represents the degree to which Wind is being high to the
North as
HighWindDirection(x) = trapezoidal (0, 250, 18, 50, 62, 70
= triangular (0, 250, 18, 50, 62)
= right (0, 250, 18, 50)
= left (0, 250, 18, 50)
13. Fuzzy OWL 2 Annotation
(a) Left-shoulder function (b) Right-shoulder function
(C) Triangular function (d) Trapezoidal function
14. Fuzzy OWL 2 Annotation
Applying the definition then the concept C mapped
CI: ∆I → [0, 1].
Implies CI is satisfiable because x ∈ ∆I,
CI(x) > 0
C can be considered as satisfiable since in KB, I determines the maximum degree of
truth that the concept C may have over all individuals, x ∈ ∆I.
16. Fuzzy OWL 2 Annotation
Fuzzy Annotation property
Fuzzy annotation for the speed of wind direction.
17. Conclusion
The expressiveness of fuzzy OWL 2 knowledge has been achieved based on language
representation using meteorological data.
The proposed method suggests the use of fuzzy OWL 2 for solving the problem of uncertainty in
meteorological data.
The presence of fuzzy OWL 2 power will reduce the ambiguity of information in the knowledge
base.
Finally suggests the use of ontology editor and reasoners in providing the error-free annotation.
18. Reference
[1] F. Bader et al. (editors): The description Logic Handbook (Theory, Implementation and Applications), Cambridge
University Press, 2003.
[2] Bobilloa, F., Straccia, U.: Fuzzy Description Logics with General t-norms and Data Types, Fuzzy Sets and
Systems, vol. 160, pp.3382–3402 (2009).
[3] C. A. Yeung and H. Leung, "A formal model of ontology for handling fuzzy membership and typicality of
instances", The computer journal, vol. 53, No. 3, 2010.
[4] Horrocks, I, Glimm, B., Sattler, U.: Hybrid Logics and Ontology Languages, Electronic Notes in theoretical
Computer Science, vol. 174, pp. 3—14 (2007).
[5] http://www.w3.org/2007/OWL/wiki/Primer
[7] Bobillo, F., Straccia, U.: Fuzzy Ontology Representation Using OWL 2, International Journal of Approximate
Reasoning, vol. 52, 1073-1094 (2011)
[8] Russell, S. J., Norvig, P.: Artificial Intelligence: A Modern Approach (2nd eds) Pearson Education, New Jersey
(2010)
[9] Fuzzy ontology plug-in (fuzzyDL 1.1). Available at: http://nemis.isti.cnr.it/~straccia/software/fuzzyDL/fuzzyDL.html
19. Universiti Teknologi PETRONAS
Department of Computer & Information Sciences
Seri Iskandar, 31750 Tronoh, Perak, Malaysia
Thank you!