A hybrid approach for standardized Dictionary-based knowledge extraction for Arabic morpho-semantic retrieval

Abstract

We propose in this paper to exploit Arabic dictionaries to enhance Arabic Information Retrieval (IR). We use standardized LMF dictionaries. We first put forward to mine such dictionaries and to represent them into graph-based representation. This graph will also be mined with a hybrid approach that combines both linguistic and statistical techniques to extract useful knowledge for IR. We study how extracted knowledge from such resource and added to the initial queries can attentively affect the retrieval process and results. Several query expansion strategies are carried based on morphological, semantic and morpho-semantic queries terms relations.