Exploratory search with semantic transformations using collaborative knowledge bases

  • Authors:
  • Yegin Genc

  • Affiliations:
  • Stevens Institute of Technology, Hoboken, NJ, USA

  • Venue:
  • Proceedings of the 7th ACM international conference on Web search and data mining
  • Year:
  • 2014

Quantified Score

Hi-index 0.00

Visualization

Abstract

Sometimes we search for simple facts. Other times we search for relationships between concepts. While existing information retrieval systems work well for simple searches, they are less satisfying for complex inquiries because of the ill-structured nature of many searches and the cognitive load involved in the search process. Search can be improved by leveraging the network of concepts that are maintained by collaborative knowledge bases such as Wikipedia. By treating exploratory search inquires as networks of concepts -- and then mapping documents to these concepts, exploratory search performance can be improved. This method is applied to an exploratory search task: given a journal abstract, abstracts are ranked based their relevancy to the seed abstract. The results show comparable relevancy scores to state of the art techniques while at the same time providing better diversity.