Authors:
Laroussi Merhbene
1
;
Anis Zouaghi
1
and
Mounir Zrigui
2
Affiliations:
1
UTIC , Higher School of Techniques Sciences of Tunis, Tunisia
;
2
Faculy of sciences of Monastir, Tunisia
Keyword(s):
Arabic ambiguous words, LSA, Harman, Okapi, Croft, Lesk algorithm, Signatures and stemmer.
Related
Ontology
Subjects/Areas/Topics:
Applications
;
Artificial Intelligence
;
Knowledge Engineering and Ontology Development
;
Knowledge-Based Systems
;
Natural Language Processing
;
Pattern Recognition
;
Symbolic Systems
Abstract:
In this paper we propose an hybrid system of Arabic words disambiguation. To achieve this goal we use the methods employed in the domain of information retrieval: Latent semantic analysis, Harman, Croft, Okapi, combined to the lesk algorithm. These methods are used to estimate the most relevant sense of the ambiguous word. This estimation is based on the calculation of the proximity between the current context (Context of the ambiguous word), and the different contexts of use of each meaning of the word. The Lesk algorithm is used to assign the correct sense of those proposed by the LSA, Harman, Croft and Okapi. The results found by the proposed system are satisfactory, we obtained a rate of disambiguation equal to 76%.