Comparing Taxonomies for Organising Collections of Documents

Proceedings of COLING 2012: Technical Papers, pages 879–894, COLING 2012, Mumbai, December 2012. Comparing taxonomies for organising collections of documents Samuel Fernando1 Mark Hall1,2 Eneko Agirre3 Aitor Soroa3 Paul Clough2 Mark Stevenson1 (1) Department of Computer Science, Regent Court, 211 Portobello, Sheffield, S1 4DP, UK (2) Information School, Regent Court, 211 Portobello, Sheffield, S1 4DP, UK (3) Universidad del País Vasco, Barrio Sarriena s/n, 48940 Leioa, Bizkaia {s.fernando,m.mhall,p.d.clough,m.stevenson}@sheffield.ac.uk {e.agirre, a.soroa}@ehu.es ABSTRACT There is a demand for taxonomies to organise large collections of documents into categories for browsing and exploration. This paper examines four existing taxonomies that have been manually created, along with two methods for deriving taxonomies automatically from data items. We use these taxonomies to organise items from a large online cultural heritage collection. We then present two human evaluations of the taxonomies. The first measures the cohesion of the taxonomies to determine how well they group together similar items under the same concept node. The second analyses the concept relations in the taxonomies. The results show that the manual taxonomies have high quality well defined relations. However the novel automatic method is found to generate very high cohesion. TITLE AND ABSTRACT IN BASQUE Dokumentu bildumak antolatzeko taxonomien arteko alderaketa Dokumentu bildumak kategorietan sailkatzea oso erabilgarria da, dokumentuak arakatzeko eta aztertzeko aukera berriak eskaintzen duen heinean. Hori horrela izanik, dokumentu bilduma handiak sailkatzeko taxonomien behar handia dago. Artikulu honetan eskuz sortutako lau taxonomia aztertzen dira, taxonomiak automatikoki sortzen dituzten bi metodorekin batera. Taxonomia hauek ondare kulturaleko bilduma handiak antolatzeko erabili ditugu. Taxonomien ebaluazioa egin dugu galdetegietan oinarritutako bi metodo erabiliaz. Lehenbizikoak tax- onomiaren kohesioa neurtzen du, hau da, antzeko itemak kontzeptu beraren azpian zein ondo taldekatzen diren. Bigarrenak taxonomiako kontzeptuen arteko erlazioak aztertzen ditu. Emaitzek erakusten dute eskuzko taxonomien erlazioen kalitatea, baina metodo automatiko berri batek lortzen du kohesio handiena. KEYWORDS: Semantic network, taxonomy, hierarchy, Wikipedia, WordNet, ontology. KEYWORDS IN BASQUE: Sare semantiko, taxonomia, hierarkia, Wikipedia, WordNet, Ontolo- gia. 879 1 Introduction With increasingly large sets of diverse collections of documents available online a key challenge is organising and presenting these items effectively for information access. To enable the navigation and exploration of collections, content providers typically provide users with free- text search functionalities, along with some form of browsable subject categories or taxonomy, also useful in organising documents. Providing multiple mechanisms for accessing documents enables users to conduct various modes of information seeking activity, from locating specific documents to more exploratory forms of searching and browsing behaviour (Hearst, 2009; Marchionini, 2006; Wilson et al., 2010). In this paper we focus on evaluating different taxonomies that could be used to organise and navigate content from Europeana1, an online cultural heritage collection. This collection comprises many subcollections taken from different providers, and thus contains a very diverse set of cultural heritage items2. Some of the subcollections are linked to bespoke taxonomies; however, many are not. This therefore represents a very challenging dataset to organise in a consistent and uniform manner. We focus on two main approaches for organising content: the …

Content truncated to 500 words