Visual Search in Digital Libraries and the Usage of External Terms

This paper focuses on the application of visual search interfaces in the context of digital libraries. The main objective is to represent a simplified and intuitive interactive approach for retrieving similar publications based on a preselected one. This would enable the scholar to perform more detailed research with the reduced mental workload, in comparison to traditional keyword-based search. The proposed approach, in an innate and conceptual manner, makes possible the application of suggested terms from other external resources. Accordingly, the set of terms can be extended with synonyms, narrowed, broadened or closely related terms. Such suggestions may result from a simple language thesaurus, any SKOS modelling scheme, and the deployment of word embedding approaches, such as word2vec. To provide a better picture of why a particular publication is presented in the results list, the matched terms are colored.