“Computing with Words”-Based Concept Retrieval

Abstract

Concept retrieval aims to extract documents that are semantically similar to the query. In view of the communication gap between the user and the system at the interface level in keyword-based search systems, user input in Natural Language (NL) is preferable. However, due to the inherent limitations of the natural languages, processing NL queries is challenging. Existing information retrieval system with conceptual search capabilities, formulate a keyword-based query out of the NL query which still might not fully reflect the concept expressed by the user. In such a scenario, the application of Computing with Words (CWW) computation is natural. In the literature, CWW techniques are applied in various processes of IR systems. In this paper, we propose a novel CWW-based paradigm for an extended idea of concept retrieval systems which are capable of returning a set of objects which satisfy the concept/constraint expressed in the user input NL description. The paper demonstrates the applicability of CWW computation for realizing concept retrieval task and thus highlights a new domain for further research.