Mining user queries with information extraction methods and linked data

Abstract

Purpose Advanced usage of Web Analytics tools allows to capture the content of user queries. Despite their relevant nature, the manual analysis of large volumes of user queries is problematic. This paper demonstrates the potential of using information extraction techniques and Linked Data to gather a better understanding of the nature of user queries in an automated manner.Design/methodology/approach The paper presents a large-scale case-study conducted at the Royal Library of Belgium consisting of a data set of 83 854 queries resulting from 29 812 visits over a 12 month period of the historical newspapers platform BelgicaPress. By making use of information extraction methods, knowledge bases and various authority files, this paper presents the possibilities and limits to identify what percentage of end users are looking for person and place names.Findings Based on a quantitative assessment, our method can successfully identify the majority of person and place names from user queries. Due to the specific character of user queries and the nature of the knowledge bases used, a limited amount of queries remained too ambiguous to be treated in an automated manner.Originality/value This paper demonstrates in an empirical manner both the possibilities and limits of gaining more insights from user queries extracted from a Web Analytics tool and analysed with the help of information extraction tools and knowledge bases. Methods and tools used are generalisable and can be reused by other collection holders.

Linked Data and Cultural Heritage

“…Universities were the most highly represented organisation type in this study. Nine National GLAM organisations were involved in the collected studies: the Biblioteca Nacional de España [21], the British National Archives [22], the National Audiovisual Archive of Finland [23], the National Library of Finland [24], the National Library of Latvia [25], the Royal Library of Belgium, [26], Bibliotheque Nationale de France [27], the National Library of Israel [28], and the National Library of Ireland [29]. This was the second most frequently represented organisation type.…”

“…This theme is closely related to both Research and Education, but covers the remaining motivating factors derived from specific needs. In some cases, these needs were taken from user needs analysis [22,26] and some were directed towards improving user experience more generally [32].…”

Named Entity Recognition and Classification on Historical Documents: A Survey

“…31 Next, in the context of the HIPE-2020 shared task, fastText word embeddings and flair contextualised string embeddings were made available as auxiliary resources for participants. 32 They were trained on newspaper materials in French, German and English, and cover roughly 18C-21C (full details in [55] and [52]). Similarly, Hosseini et al [87] published a collection of static (word2vec, fastText) and contextualised embeddings (flair) trained on the Microsoft British Library (MBL) corpus.…”

“…Figure 1), primary needs also revolve around retrieving documents and information, and NE processing is of similar importance [35]. There are less query logs over historical collections than for the contemporary web, but several studies demonstrate how prevalent entity names are in humanities users' searches: 80% of search queries on the national library of France's portal Gallica contain a proper name [33], and geographical and person names dominate the searches of various digital libraries, be they of artworks, domain-specific historical documents, historical newspapers, or broadcasts [14,32,92]. Along the same line, several user studies emphasise the role of entities in various phases of the information-seeking workflow of historians [47,71], now also reflected in the 'must-have' of exploration interfaces, e.g.…”

Large Scale Analysis of Semantic and Temporal Aspects in Cultural Heritage Collection's Search

“…However, as users get more experienced, they reduce the number of operations in their sessions. Another work [4] presented the results of a large-scale case-study at the Royal Library of Belgium based on a data set of 83k queries from 29k visits over a year long period of the historical newspapers platform BelgicaPress (associated with the State Archives of Belgium). The authors investigated the application of simple text mining methods such as query clustering in cultural heritage settings.…”