Ontology Alignment Revisited
Total Page：16
File Type：pdf, Size：1020Kb
Semantic Web 1 (0) 1–5 1 IOS Press 1 1 2 2 3 3 4 Ontology Alignment Revisited: A 4 5 5 6 Bibliometric Narrative 6 7 7 8 Majid Mohammadi a,*, Amir Ebrahimi Fard a 8 9 a Delft University of Technology, The Netherlands 9 10 10 11 11 12 12 13 13 14 Abstract. Ontology alignment is an important problem in the Semantic Web with diverse applications in various disciplines. 14 15 This paper delineates this vital field of study by analyzing a core set of research outputs from the domain. In this regard, the 15 16 related publication records are extracted for the period of 2001 to 2018 by using a proper inquiry on the well-known database 16 Scopus. The article details the evolution and progress of ontology alignment since its genesis by conducting two classes of 17 17 analyses, namely, semantic and structural, on the retrieved publication records from Scopus. Semantic analysis entails the overall 18 18 discovery of concepts, notions, and research lines flowing underneath ontology alignment, while the structural analysis provides 19 a meta-level overview of the field by probing into the collaboration network and citation analysis in author and country levels. 19 20 In addition to these analyses, the paper discusses the limitations and puts forward lines for the further progress of ontology 20 21 alignment. 21 22 22 23 Keywords: ontology alignment, bibliometrics, scientometric 23 24 24 25 25 26 26 27 1. Introduction In this regard, a pre-processing strategy is required to 27 28 align the ontologies of these heterogeneous informa- 28 29 The Semantic Web is an extension of the World tion systems, after which they can interact with each 29 30 Wide Web which aims to provide metadata for ma- other. This strategy is called ontology alignment (also 30 31 chines so that they can further understand and construe called ontology mapping and ontology matching). 31 32 the published information and data on the Web. The The necessity of having a tool to automatically align 32 33 consequence of having such metadata is that comput- two different ontologies was recognized in the late 33 34 ers can also make reasonable interpretations, ideally 90s and early 2000 [73–75, 98]. Since the heterogene- 34 35 identical to how humans process and understand the ity problem was quite epidemic and was a stumbling 35 36 information. The main key to providing such informa- block in the way of interoperability, it soon found its 36 37 tion for computers is ontologies, with which one can way in many applications such as agent communica- 37 38 model the underlying objects of a domain along with tion in agent-based modeling [98], ontology merging 38 39 their interrelations. The design of ontologies is thor- [74, 75], data integration [80], business-to-business e- 39 40 oughly subjective and is primarily reliant on the vi- commerce [76], ontology development and visioning 40 41 sion of the creator. Since humans potentially consider [16], database evolution [17, 58], web service discov- 41 42 different aspects of a domain, or they might use dif- ery [23], to name just a few. Due to the diverse ap- 42 43 ferent terminology for similar concepts, the ontologies plications of ontology alignment, many research stud- 43 44 are not uniquely defined and are distinct from each ies have been dedicated to resolving the heterogene- 44 45 other, even those from one particular domain. If infor- ity among information systems, and new problems are 45 46 mation systems using these ontologies are assumed to modeled to be solved by alignment techniques. 46 47 work independently, the difference in ontologies does In recent two decades, tremendous efforts have been 47 48 not cause any issue. However, the issue emerges when taken to further improve the field of ontology align- 48 49 these systems want to interoperate and exchange data. ment [26, 45, 67–70, 77]. As a result of such efforts, 49 50 there are several valuable materials available for ontol- 50 51 *Corresponding author. E-mail: [email protected]. ogy alignment researchers and others who want to uti- 51 1570-0844/0-1900/$35.00 c 0 – IOS Press and the authors. All rights reserved 2 M. Mohammadi and A. Ebrahimi Fard / Ontology Alignment Revisited: A Bibliometric Narrative 1 lize or get acquainted with alignment techniques. For a parameter-free measure that quantifies the extent to 1 2 instance, there is a book containing the fundamentals which a specific paper can be considered an SB. 2 3 of ontology alignment and recent advances made in The other line of studies in bibliometrics calibrates 3 4 this field [26]. Further, there are some useful reviews research activities to provide insights regarding dy- 4 5 and surveys [77, 94] and some other stating the cur- namic and vital influential factors behind scientific re- 5 6 rent state and future challenges of ontology alignment search. Citation analysis [38, 88], co-authorship analy- 6 7 [89]. Although these materials are essential and help sis [35, 85, 88], and co-occurrence word analysis [15] 7 8 researchers get familiar with the notions of ontology are prevalent in this application domain of bibliomet- 8 9 alignment, they do not provide an overview of the field. rics. For instance, in the study carried out by Bromham 9 10 Such an overview is not only critical to those who want et al. [8] on The Australian Research Council’s grant 10 11 to get familiar with the domain, but it is also vital for proposal data, they studied the relationship between 11 12 researchers in this field to track its progress and find research interdisciplinarity and the chance of winning 12 13 new ways to further improve it. In addition, such stud- grants and discovered the higher the degree of interdis- 13 14 ies allow science policymakers to get a comprehensive ciplinarity, the lower the probability of being funded. 14 15 picture of the field. This enables them to assess the Also, in another research focusing on collaboration 15 16 scientific advancement of this subject domain. Assess- in the field of Genomics, Petersen et al. [79] discov- 16 17 ment is a vital element in decision-making as it gives a ered cross-disciplinary research draw more attention 17 18 realistic picture of the current situation and prevent us and get more citations correspondingly. One type of 18 19 from over- or under-estimation. Therefore, to be sure research that falls into the same line of research is the 19 20 the right decisions are made for this field to proceed in bibliometric analysis of scientific fields. For instance, 20 21 21 the right path, assessment is essential. In this regard, Frank et al. [32] studied the bibliometric evolution of 22 AI research and its related fields since 1950. 22 this article utilizes the bibliometric analysis to provide 23 On the other side of the spectrum, bibliometrics is 23 a bird’s-eye view to the interdisciplinary domain of on- 24 utilized to address much broader goals. Some studies 24 tology alignment. 25 use bibliometrics data or analysis for answering ques- 25 Bibliometrics is a quantitative approach to study sci- 26 tions which are not for the purpose of scientific ac- 26 entific activities. At its most fundamental level, biblio- 27 tivities evaluation. This is a very recent approach to- 27 metrics aims to unveil the latent dynamics of scientific 28 ward bibliometrics, which can provide an opportunity 28 research and analyze its key influential factors. Con- 29 for other disciplines to benefit the tools and techniques 29 tent, citation, and collaboration analyses are among 30 developed in this field. For instance, Candia et al. [10] 30 the commonly-practiced techniques using bibliomet- 31 studied the problem of collective memory decay us- 31 rics analysis. In this regard, many research fields use 32 ing multiple datasets including American Physical So- 32 these techniques to delineate the importance of their 33 ciety (APS) papers and the United States Patent and 33 34 fields, the impact of the lead researchers, or gauge the Trademark Office (USPTO) patents. In the other work, 34 35 impact of a particular research output [61]. In recent Guimera et al. [36] studied the self-assembly of cre- 35 36 years, the bibliometric analysis has drawn a lot of at- ative teams in the collaboration network using empiri- 36 37 tention and is applied to major bibliometrics data en- cal study over a bibliometric dataset constitutes of 50 37 38 gines. Bibliometrics covers a broad spectrum of ap- years records of recognized journals in social psychol- 38 39 plication domains [31]. Some of the studies in biblio- ogy, ecology, economics, and astronomy. In another 39 40 metrics have more methodological orientations and try research, Liu et al. [60] studied the phenomenon of a 40 41 to scrutinize the existing bibliometric measures, e.g., hot streak for individuals career. They conducted this 41 42 citation and impact factor, or to come up with new study by combining over 20,000 researcher profiles in 42 43 ones. For instance, Chorus et al. [12] defined a met- Google Scholar and Web of Science. Ebrahimi Fard 43 44 ric for self-citation and studied trends in impact factor et al. [21] also used bibliometric analysis to study the 44 45 biased self-citations of scholarly journals. In the other readiness of academia amid a war with the diffusion of 45 46 research, Thelwall et al. [93] made a comparison for fake-news in social media. 46 47 11 altmetrics in Web of Science to understand the re- This article brings forth a bibliometric approach 47 48 lationship between real citations and alternative met- to analyze the growth and advancement of ontology 48 49 rics in social media.