International Journal of Computer Applications |
Foundation of Computer Science (FCS), NY, USA |
Volume 63 - Number 14 |
Year of Publication: 2013 |
Authors: Abdessalem Kelaiaia, Hayet Farida Merouani |
10.5120/10536-5529 |
Abdessalem Kelaiaia, Hayet Farida Merouani . Influence of stemming on Clustering of Arabic texts: Comparative Study in Document Retrieval. International Journal of Computer Applications. 63, 14 ( February 2013), 36-41. DOI=10.5120/10536-5529
Initially, this paper, sets out to study the influence of stemming on the quality of the Arabic text clustering, and then describes the testing the application of an approach based on this clustering to improve Document Retrieval (DR). A classical local document system generally, employs statistical methods for calculating the similarity between the introduced query and each document in the target collection to finally provide an ordered list of documents (hit list). In the present approach, the collection is submitted to the clustering process, and then the list of documents returned is constructed from formed clusters based on the nearest representative among the representatives of clusters compared to the user's query. The choice of the Arabic language is motivated by its very particular morpho-syntactic characteristics.