Abstract
Exploring publications in academia usually entails browsing a collection of published literature. In the digital space, collections typically come in the form of primitive bibliographic databases or full featured digital libraries. Bibliographic databases contain an organized collection of references to published literature, whereas digital libraries host publications in full-text. Although lacking full-text content of a publication, bibliographies contain rich metadata such as publication titles, venues or contributing authors that are adequate enough for finding publications of interest. One example of a bibliographic database is DBLP, which provides publication records in the computer science discipline. DBLP supports researchers in publication exploration by providing interfaces for finding specific bibliographic records or collections of records by a specific author. However, it does not provide a means for exploring families of similar publications by topic areas. In this project, an application tool named DBLPminer is developed that extends upon the DBLP dataset. Addressing the topic limitations of DBLP, DBLPminer provides an interface for linking and accessing DBLP records to topic areas. Using this tool, researchers may organize DBLP records by topic categories and subsequently explore publications by topic categories or find associated topics for a DBLP record. At its foundation, the application indexes DBLP records by topic categories under the taxonomy of ACM 2012 Computing Classification System. Indexed DBLP publication records are made accessible via a prototype web application interface. Lastly, analysis is made on these datasets and the indexing algorithm's performance. There exist other tools like DBLPminer that aid researchers in exploring publications through bibliographic data. Some tools take a social network approach such as Arnetminer and ResearchGate, allowing researchers to connect with one another to share and access publications. Tools such as Scholarometer and Google Scholar are more search oriented with interfaces for finding publications of interest. Although similar, DBLPminer differs from tools such as Arnetminer, ResearchGate, and Scholarometer because it focuses on the publication relationships by topic instead of authorship and references. Another difference between DBLPminer and ResearchGate, Scholarometer, and Google Scholar is that it focuses on publications specific to the computer science community.