LabJam (Internship)

LabJam (Internship) is a DCU School of Computing-funded internship, 4 months in duration (November 2010 - March 2011), with 1 academic partner (Dublin City University).

DCU is developing an experimental research management system, called LabJam, that enables researchers and their managers to capture and collate disparate forms of R&D output in such a way as to facilitate enhanced integrated analysis and cohesive reporting of impact.

The first phase of this project involves developing the base platform for evaluation. The base platform is based on a set of services, supported by a public facing API, and a user-controlled set of fine-grained access control policies that allow the balance between publication and IP protection to be systematically managed.

The second phase evaluates three approaches to assessing research impact. The three approaches we endeavour to evaluate and exploit in this task include: Expert search, bibliometric analysis, and user-centric recommendations. Expert search provides researchers with a way to quickly asses the leading contributors in a particular field, to survey the emergence of new and important researchers and to map relations between fields. Expert search is based on multiple aspects of publication, including its abstract, categories (e.g. the ACM classification), body of text, and citations. Bibliometric analysis provides a set of tools to analyse the impact of research publications by indexing citations tracked across portals such as IEEE, ACM and Elsevier. Similar to expert search, bibliometric analysis is based on the relations between citations. For example, any two papers cited within the same paper can be assumed to be related as well as two papers citing the same paper. In addition, traditional bibliometric scores (such as the H-score) can be computed to identify the long-term impact of publications on research using citation counts. Both components (expert search and bibliometric analysis) are based on a semantic analysis of textual information to extract citations and related metadata. The semantic analysis is then coupled with social network analysis to develop a more complete vista of a discipline and related fields. User-centric recommendations can then be provided based on a combination of the above analytic techniques, coupled with individual researcher profiles.

Please contact

• Dr. Johannes Leveling