Semantic CiteSeerx

Semantic CiteSeerx will enhance CiteSeerx with intelligent functionality by mining CiteSeerx’s extensive collection of research literature to provide users with knowledge they can use to improve and expedite their research from the planning stage to the publication of research results. The efforts here at University of Arkansas mainly focus on designing and developing new personalization features for CiteSeer. Our approach to personalization is based on conceptual user profiles. Our previous experience with KeyConcept proved that conceptual profiles are an efficient way to represent user interests.

Our goals are:

  • Study new models for recommendation engine including publication venues recommendation, suggesting collaborators, finding experts in a domain, summarizing state of the art of a research field, recommending papers to reviewers, etc.
  • Design and develop publication venues recommendation using author publication history and social network analysis
  • Design and develop a system for tracking user actions and browsing history. Adapting user conceptual profiles with new extracted information.
  • Provide personalized browsing interfaces for the CiteSeerx users.
  • Develop an ontology-based hierarchy browsing system to help users browse documents by categories easier.
  • Design and develop a Conceptual Recommender System using the conceptual user profiles.
  • Using machine learning techniques to classify and cluster documents in the CiteSeerx database and new ones that are regularly added to the collection.

Project Information:

Project Name

Collaborative Research: CI-ADDO-EN: Semantic CiteseerX

Award Number

0958123

Start Date

July 1, 2010

Investigators

Susan Gauch (Principal Investigator)
Hiep Luong (Co-Principal Investigator)

NSF Program

EXP PROG TO STIM COMP RES, COMPUTING RES INFRASTRUCTURE

NSF Org

CNS Division of Computer and Network Systems