A Taxonomy and Survey of Semantic Approaches for Query Expansion

Muhammad Ahsan Raza, M. Rahmah, A. Noraziah, Maruf Pasha, Urooj Pasha
2019 IEEE Access  
Conventional approaches to query expansion (QE) rely on the integration of an unstructured corpus and probabilistic rules for the extraction of candidate expansion terms. These methods do not consider search query semantics, thereby resulting in ineffective retrieval of information. The semantic approaches for QE overcome this limitation, whereby a search query is expanded with meaningful terms that accord with user information needs. This paper surveys recent approaches to semantic QE that
more » ... oy different models and strategies and leverages various knowledge structures. We organize these approaches into a taxonomy that includes linguistic methods, ontology-based methods, and mixed-mode methods. We also discuss the strengths and limitations of each type of semantic QE method. In addition, we evaluate various semantic QE approaches in terms of knowledge structure utilization, corpus collection, baseline model adaptation, and retrieval performance. Finally, future directions in exploiting personalized social information and multiple ontologies for semantic QE are suggested. INDEX TERMS Information retrieval, morphological expansion, ontology, semantic query expansion.
doi:10.1109/access.2019.2894679 fatcat:ohugfhrpzjadtpc5zv7uwabk4i