Arabic Information Retrieval: A Relevancy Assessment Survey

The paper presents a research in Arabic Information Retrieval (IR). It surveys the impact of statistical and morphological analysis of Arabic text in improving Arabic IR relevancy. We investigated the contributions of Stemming, Indexing, Query Expansion, Text Summarization (TS), Text Translation, and Named Entity Recognition (NER) in enhancing the relevancy of Arabic IR. Our survey emphasizing on the quantitative relevancy measurements provided in the surveyed publications. The paper shows that the researchers achieved significant enhancements especially in building accurate stemmers, with accuracy reaches 97%, and in measuring the impact of different indexing strategies. Query expansion and Text Translation showed positive relevancy effect. However, other tasks such as NER and TS still need more research to realize their impact on Arabic IR.