![](bg6.png)
for each pair of words detected, the relevance of the
terms and the semantic relationship between them.
Two values are possible:
The value "valid": if the relationship is deemed
valid and the two words linked by this
relationship;
The value "false": when the relationship is
wrong (in this case, the terms are often invalid).
Obtained results are shown in Table 4.
Table 4: Evaluation results of morpho-lexical patterns.
Following the evaluation of built grammars, we
obtain values of recall, precision and F-measure,
respectively: 78%, 85% and 79%. Although the
results are encouraging, there are relationships that
are not detected by the defined patterns. By
analyzing non-detecting pairs of words, we find that
this problem is mainly due to two reasons:
The lack of patterns that cannot recognize the
relationship: In fact, in some cases the patterns,
concerning some phrases, are not defined;
The non-detection of compound nouns: using
NooJ morphological analyzer, we notice that it
is unable to detect compound nouns, which
prevents from applying some patterns to the
sentences.
Moreover, in some cases we notice that the
patterns recognize erroneous relationships. This can
be explained by two reasons:
The existence of some ambiguous patterns;
Morphological analysis using NooJ platform
produce several morphological categories: in
some cases NooJ return more than one
morphological category of a word, which allow
to us apply several patterns on one sentence and
detect more than one semantic relation between
two terms.
6 CONCLUSIONS AND FUTURE
WORK
In this paper, we propose a method to define
morpho-lexical patterns that is useful for extracting
semantic relations between arabic nouns. This
method is based on a corpus of 2050 Arabic
Wikipedia articles. We have detailed the various
steps of the proposed method. As a result, the
defined patterns are implemented using NooJ
platform in order to automatically identify the pairs
of words and the semantic relations between these
specific terms. Nevertheless, our method uses a
minimum of knowledge, based primarily on
morpho-lexical knowledge, obtained results are very
encouraging (i.e. 78% recall, 85% precision and
79% F-measure) which proves the importance of
morpho-lexical patterns in the detection of semantic
relations for Arabic language.
As future perspectives, we plan to resolve
ambiguous patterns. In addition, we intend to
propose a method for detecting compound nouns.
Finally, we consider applying these patterns to build
a lexical ontology for Arabic.
REFERENCES
Baccar, F., Gargouri, B., Ben Hamadou, A., 2012.
Towards Generation of Domain Ontology from LMF
Standardized Dictionaries. In Proceedings of the 22nd
International Conference on Software Engineering &
Knowledge Engineering (SEKE).
Ben Mustapha, N., Aufaure, M. A., Baazaoui Zghal, H.,
2006. Vers une approche de construction de
composants ontologiques pour le Web sémantique –
synthèse et discussion. In Troisième atelier Fouille de
Données Complexes dans un processus d’extraction
des connaissances.
Arara, A., Ben Slimane, D., 2002. Towards ontologies
building: a terminology based approach, In VLDB
Journal - Special Issue on Semantic Web.
Gurber, T. R., 1999. Toward Principles for the Design of
ontologiesused for Knowledge Sharing. In
International Journal of Human-Computer Studies.
Vol. 43 , pp. 907-928.
Borst, W., 1997. Construction of Engineering Ontologies
for Knowledge Sharing and Reuse. Ph.D.
Dissertation.
Studer, R., Benjamins, V. R., Fensel, D., 1998.
Knowledge Engineering: Principles and Methods. In
Data Knowl. Eng., vol. 25, pp. 161–197.
Gómez-Pérez, A., Fernández-López, M., Corcho, O.,
2004. Ontological engineering: with examples from
the areas of knowledge Management, e-Commerce
and the semantic Web.
George, A. M., Beckwith, R., Fellbaum, C., Gross, D.,
Miller K., 1993. Introduction to wordnet: An On-line
Lexical Database. In International Journal of
Lexicography.
Vossen, P., 1998. Eurowordnet: A Multilingual Database
with Lexical Semantic Networks. Livre.
Laignel, M., Kamel, M., Aussenac-Gilles, N., 2011.
Enrichir la notion de patron par la prise en compte de
la structure textuelle - Application à la construction
d’ontologie. In TALN 2011 (Traitement automatique
des langues naturelles).
Black, W., Elkateb, A., Rodriguez, H., Alkhalifa, M.,
Vossen, P., Pease, A., Fellbaum, C., 2006. Introducing
the Arabic wordnet Project.
Towards an Arabic Ontology - Defining Morpho-lexical Patterns for Semantic Relation Extraction
347