Abstract
This paper introduces an approach that combines a set of semantic techniques to enhance traditional information retrieval for the Arabic language. The approach uses a set of linguistic techniques to deal with the Web content. It expands first the initial user's query with additional related words by using lexical semantic relations (synonyms) covered in Arabic WordNet thesaurus and by domain specific words and abbreviations collected from the corpus. Second, it makes use of linguistic methods to match semantically the query objective with the text content in the search results. The paper presents the approach used and also the implementation of the system. Experimental results show that semantic retrieval outperforms the traditional information retrieval which is based only on query keywords matching.