Professional Networking and Expertise Mining for Research Collaboration
Profiles Research Networking Software is an NIH-funded open source tool to speed the process of finding researchers with specific areas of expertise for collaboration and professional networking. Profiles RNS imports and analyzes "white pages" information, publications, and other data sources to create and maintain a complete searchable library of web-based electronic CV's. Built-in network analysis and data visualization tools allow administrators to generate research portfolios of their institution, discover connections between parts of their organization, and understand what factors influence collaboration.
Profiles RNS self-populates a database of publication history, research interests and professional relationships for each investigator in an organization. Integrated visualization and search tools make Profiles RNS easy to use, and its customizable look-and-feel allow Profiles RNS to be integrated into an existing website or set up as a stand-alone site. Profiles RNS data is also accessible through an API to power other applications.
Profiles RNS pioneered the concept of "passive" and "active" networking, which not only enables the website to be a useful and exciting tool on day one, but also allows users to expand its content with information about social networks that only they know.
More Than Simple Search Results
Profiles RNS provides much more useful information than typical directory listings or ordinary literature searches. Algorithms analyze publication data to define a researcher's professional interests with a set of prioritized keywords.
The factors used to rank and weight the significance of a specific keyword as a useful descriptor of a researcher include:
The Profiles RNS Author Disambiguation Engine
Profiles RNS uses sophisticated multi-factorial matching algorithms to build a publication history automatically for each researcher in an institution. This "Disambiguation Engine" self-populates the individual researcher overviews in Profiles RNS, and identifies the specific keywords that characterize each researcher.
Using identity information from a managed data source such as a Human Resources database, the Disambiguation Engine extracts citations from PubMed and assigns publications to specific individuals. The Disambiguation Engine uses a number of factors to build each publication history, including:
Users (or their proxies) can add any missing publications by doing a PubMed search from within Profiles RNS or manually entering publications that do not exist in PubMed. The Disambiguation Engine learns from these changes to improve the results of the next literature analysis and update.
Profiles RNS creates a "career snapshot" that combines directory information, user-contributed content, and publications that are extracted from PubMed by the Profiles RNS Disambiguation Engine. On the right side-bar, Profiles RNS automatically identifies networks of related people and concepts ("passive" networking). On the left, users can create "active networks" with specific colleagues.
User Managed Content and Privacy Controls
Each researcher at an institution has control over her or his information. While contact details and other directory information are managed through the original source system at the institution, each user can select which sections of their overview page is displayed or hidden. In addition to editing their publication lists, users can add a photograph, short narrative summary, and awards to their profiles.
To make managing researcher information easier, Profiles RNS allows proxies to be designated for each user. Proxy access is configurable, controlling the ability to edit or show/hide specific categories of information.
Federated Search and Standards-Based APIs
A significant new feature in Profiles RNS is Federated Search. Instances of Profiles RNS at different institutions can "talk" to each other through federated queries, enabling networks to extend beyond local collaborations, and protecting institutional data.
With a Pilot project being launched in early 2011, users will also be able to search for potential research collaborators at participating institutions that have other professional networking tools such as VIVO and Loki.
These capabilities are possible because Profiles RNS uses standards-based web services APIs that can communicate with other computer systems through XML, RIS Citation Data Format, RDF and Linked Open Data (LOD).
Extending Functionality Through an Ontology
Profiles RNS is an ontology-based application, meaning users with administrative privileges can extend the types of data that are supported by simply describing the new entities and their relationships. For example, by stating that a person can "teach" a "course", a new content section for teaching will be added to users' profiles. By default, Profiles RNS uses the VIVO ontology. VIVO is an NIH funded project that created an RDF-based standard for sharing information about researchers. In addition to this "data" ontology, Profiles RNS also has a "presentation" ontology, which allows customization of the layout of the website and defines security rules that determine who can access what data.