Socio-Semantic Information Retrieval of Structured Arabic Texts

We propose in this paper a new socio-semantic information retrieval system denoted RISSA. The indexing units are extracted based on the hybrid possibilistic morphological disambiguation approach. Then we improve this approach by integrating socio-semantic features. We propose a new approach that takes advantage of the social and semantic diversity and the structuring of the Arabic Hadith texts. The social aspect of the Hadith texts was highlighted by presenting a new search mode by Sanad and/or Metn and/or verses of Quran and/or poems and/or section titles. Thus, the search query can be observed along several axes: (i) A social axis taking advantage of the existing social relationships in the Hadith chains of narrators (Sanad), (ii) A semantic axis taking into account the Hadith decomposition in several topics useful to identify the meaning of the texts (Metn). We firstly experiment the RISSA system using the Kunuz collection of Hadith. Then, we investigate and discuss the impact of socio-semantic features on the morphological disambiguation of the structured Hadith Arabic texts.