From Controlled Vocabularies to Ontologies: A Techno-Human Approach

Abstract

The biggest challenge search platforms face is how to lead the users towards the answers they are looking for with very few clicks. The quest for the most straightforward route to the most relevant result becomes particularly daunting when users start their journey with a broad query (1-2 word), providing no context around their intentions. This leads to voluminous and unfocused search results and leaves the user with the burden of identifying the most relevant ones. Allowing users to refine large result sets by selecting query-relevant concepts from a controlled vocabulary is a good option; however, creating a content-tailored vocabulary is a challenge in itself, especially when the vocabulary has to cover diverse domains. In this talk, I will showcase Knovel’s techno-human approach toward generating a content-centric, multi-domain vocabulary to use as refinement tool, and the subsequent strategy to extract from the corpus ontological relationships between the concepts in the vocabulary.

Keywords: Search, Information Retrieval, Machine Extraction, Content Tagging, Concept Faceting