LIBER 2021 Session #10: Search & Retrieval: Next Generation Information Services

These are the slides for the LIBER 2021 Session Search & Retrieval: Next Generation Information Services

This session will be chaired by Karine Bacher-Eyroi, Université de Pau et des Pays de l’Adour, France

• Towards modelling text mining services for digital collections: the case of Latvian Prose Counter, Anda Baklāne, Valdis Saulespurēns, National Library of Latvia
• Creating gold standards and supervising outcomes – the future role of library staff in supporting information services based on Machine Learning, Timo Borst, ZBW Leibniz Information Center for Economics, Germany
• Speed talk: Connecting Open Knowledge Components to enhance searches for researchers and topics, Arben Hajra, Tamara Pianos, ZBW Leibniz Information Center for Economics, Germany

In the first presentation, Anda Baklāne and Valdis Saulespurēns discuss a recent paper presenting a text mining demonstration tool developed at the National Library of Latvia – the Latvian Prose Counter. The Latvian Prose Counter is an interactive website that allows users to explore various quantitative parameters of 20th century Latvian prose fiction, including counting, analysing, and visualising various lexical and syntactical features of full texts of novels and stories, as well as analysing the statistics of publishing. The workings, uses, and development of this website will be discussed in detail.
In a second presentation, Timo Borst explores the future role of library staff in supporting information services based on Machine Learning. He discusses the challenges and corresponding skills from and for (data) librarians, that are conceived as basic for the development of many, if not all AI and ML related information services. To do so, he refers to three exemplary current in-house projects at ZBW Leibniz Information Centre for Economics: automatic indexing of scientific full texts, a chatbot for serving user requests, and automatic metadata generation by identifying the funders of a scientific work resp. publication.
In a third and final presentation, Arben Hajra presents EconBiz, a subject portal for business and economics where users can easily find a list of authors who published in a specific field according to a given thesaurus concept or predefined classification code. Furthermore, through the Linked Open Data principles, it is able to crosslink and harvest different author-related data, with the purpose to ensure a comprehensive author profile. The uses and development of this tool will be explored in the presentation.