File(s) not publicly available
Reason: Waiting on paper to be submitted
A Team Assembly Tool Based on NU Scholars VIVO endpoint
online resource
posted on 2015-08-18, 11:06 authored by Raja Lalith Sabbisetti, Anup Sawant, Harshad Gado, Noshir ContractorIn earlier work, we developed a suite of heuristics for building collaborations between researchers. These heuristics are informed by empirical studies that test theories from the social sciences regarding the formation of effective collaborations and teams. In the current work, the initial prototypes of these heuristics were ported to operate over data represented in the VIVO ontology. The C-IKNOW VIVO team recommender is a web application that allows users to assemble teams based on the NU Scholars VIVO endpoint.
Our efforts demonstrate that the architectures and programming techniques of the semantic web are well suited to the problem of building practical software tools that can be leveraged to apply to diverse sources of data.In general, much of the infrastructure required for developing useful researcher recommender systems is available today. In particular, we found that the interoperability between researcher networking systems (RNSs) from diverse institutions and vendorsoffered by the VIVO ontology is a sound basis on which to build researcher networking recommender tools.
The C-IKNOW VIVO Team Recommender aims to assemble teams based on the preferences of an individual. These preferences encompass team size, homophily and network properties of co-authorship networks. The tool also aims to support team assessment feature that would allow individuals to evaluate different potential teamsbased on their preferences.
The ‘VIVO Team Recommender’ utilizes the World Wide Web Consortium (W3C) standard SPARQL query language for real-time retrieval of semantic web data. We found the SPARQL implementation available in open source software to be robust. Further, because programmers can target only the particular data needed, performance is enhanced by reducing unnecessary network traffic. We found the learning curve and technical skills needed for SPARQL programming to be similar to that of more traditional, relational, SQL-based programming competencies which are more generally available.
Our efforts demonstrate that the architectures and programming techniques of the semantic web are well suited to the problem of building practical software tools that can be leveraged to apply to diverse sources of data.In general, much of the infrastructure required for developing useful researcher recommender systems is available today. In particular, we found that the interoperability between researcher networking systems (RNSs) from diverse institutions and vendorsoffered by the VIVO ontology is a sound basis on which to build researcher networking recommender tools.
The C-IKNOW VIVO Team Recommender aims to assemble teams based on the preferences of an individual. These preferences encompass team size, homophily and network properties of co-authorship networks. The tool also aims to support team assessment feature that would allow individuals to evaluate different potential teamsbased on their preferences.
The ‘VIVO Team Recommender’ utilizes the World Wide Web Consortium (W3C) standard SPARQL query language for real-time retrieval of semantic web data. We found the SPARQL implementation available in open source software to be robust. Further, because programmers can target only the particular data needed, performance is enhanced by reducing unnecessary network traffic. We found the learning curve and technical skills needed for SPARQL programming to be similar to that of more traditional, relational, SQL-based programming competencies which are more generally available.
