Klink-2: Integrating Multiple Web Sources to Generate Semantic Topic Networks

Abstract

Abstract. The amount of scholarly data available on the web is steadily increasing, enabling different types of analytics which can provide important insights into the research activity. In order to make sense of and explore this large-scale body of knowledge we need an accurate, comprehensive and up-todate ontology of research topics. Unfortunately, human crafted classifications do not satisfy these criteria, as they evolve too slowly and tend to be too coarsegrained. Current automated methods for generating ontologies of research areas also present a number of limitations, such as: i) they do not consider the rich amount of indirect statistical and semantic relationships, which can help to understand the relation between two topics -e.g., the fact that two research areas are associated with a similar set of venues or technologies; ii) they do not distinguish between different kinds of hierarchical relationships; and iii) they are not able to handle effectively ambiguous topics characterized by a noisy set of relationships. In this paper we present Klink-2, a novel approach which improves on our earlier work on automatic generation of semantic topic networks and addresses the aforementioned limitations by taking advantage of a variety of knowledge sources available on the web. In particular, Klink-2 analyses networks of research entities (including papers, authors, venues, and technologies) to infer three kinds of semantic relationships between topics. It also identifies ambiguous keywords (e.g., "ontology") and separates them into the appropriate distinct topics -e.g., "ontology/philosophy" vs. "ontology/semantic web". Our experimental evaluation shows that the ability of Klink-2 to integrate a high number of data sources and to generate topics with accurate contextual meaning yields significant improvements over other algorithms in terms of both precision and recall.

Forecasting the Spreading of Technologies in Research Communities

“…As a reference topic ontology, we used the Computer Science Ontology (CSO), created to represent topics in the Rexplore system [3], which is currently being trialled by Springer Nature to classify proceedings in the field of Computer Science [17], such as the well-known LNCS series. CSO was created by applying the Klink-2 algorithm [18] to the 16 million publications of our Scopus-derived dataset [3]. The Klink-2 algorithm combines semantic technologies, machine learning and knowledge from external sources (e.g., DBpedia, calls for papers, web pages) to automatically generate a fully populated ontology of research areas, which uses the Klink data model 10 .…”

“…However, as extensively discussed in previous works [3,17,18], these solutions ignore the rich network of semantic relationships between research topics and are often unable to distinguish research areas from other terms that may be used to annotate publications. Therefore, we exploit the topic ontology by associating to each paper i) all the concepts in CSO whose label is found either in the title, the abstract or the keyword set, as well as ii) all skos:broaderGeneric and iii) all relatedEquivalent areas of the initial set of topics extracted from the scholarly dataset.…”

TechMiner: Extracting Technologies from Academic Publications

“…It takes as input the IDs, the titles and the abstracts of a number of research papers in the Scopus dataset 6 and a variety of knowledge bases (DBpedia [12], WordNet [15], the Klink-2 Computer Science ontology [16], and others) and returns an OWL ontology describing a number of technologies and their related research entities. These include: 1) the authors who most published on it, 2) related research areas, 3) the publications in which they appear, and, optionally, 4) the team of authors who introduced the technology and 5) the URI of the related DBpedia entity.…”

“…The Klink-2 Computer Science Ontology (CSO) is a very large ontology of Computer Science that was created by running the Klink-2 algorithm [16] on about 16 million publications in the field of Computer Science extracted from the Scopus repository. The Klink-2 algorithm combines semantic technologies, machine learning and external sources to generate a fully populated ontology of research areas.…”

“…Our intention was not to create 'yet another ontology' of the scholarly domain, but to craft a simple scheme for representing our output. For this reason we reused concepts and relationships from a number of well-known scholarly ontologies (including FABIO [22], FOAF 16 , CITO, SKOS, SRO 17 , FRBR 18 ) and introduced new entities and properties only when necessary. The main classes of the TechMiner OWL ontology are Technology, foaf:Person, to represent the researchers associated to the technology, Topic (equivalent to frbr:concept and skos:concept) and Category, representing the category of the technology (e.g., application, format, language).…”