Building Custom Term Suggestion Web Services with OAI-Harvested Open Data

TL;DR

This paper presents a method to develop custom web services for term suggestion in scientific digital libraries, leveraging open data and semantic mapping to improve search term relevance and usability.

Contribution

It introduces a framework for building domain-specific search-term recommenders using open data sources like OAI-harvested content.

Findings

Enhanced term suggestion accuracy

Improved user search experience

Potential for increased data consistency

Abstract

The problem that the same information need can be expressed in a variety of ways is especially true for scientific literature. Each scientific discipline has its own domain-specific language and vocabulary. This language is coded into documentary tools like thesauri or classifications that are used to document and describe scientific documents. When we think of information retrieval as "fundamentally a linguistic process" (Blair, 2003) users have to be aware of the most relevant search terms - which are the controlled thesauri terms the documents are described with. This can be achieved with so-called search-term-recommenders (STR) that map free search terms of a lay user to controlled vocabulary terms which can then be used as a term suggestion or to do an automatic query expansion (Hienert, Schaer, Schaible, & Mayr, 2011). State-of-the-art repository software systems like DSpace or EPrints already offer some kind of term suggestion features in search or input forms but these implementations only work as simple auto completion mechanisms that don't incorporate any kind of semantic mapping. Such software systems would gain a lot in terms of usability and data consistency if tools like the proposed domain-specific STRs would be freely available. We aim to implement a rich toolbox of web services (like the mentioned domain-specific STRs) to support users and providers of online Digital Library (DL) or repository systems.