Conflict-aware historical data fusion

  • Authors:
  • Vladimir Zadorozhny;Ying-Feng Hsu

  • Affiliations:
  • Graduate Program of Information Science and Technology, University of Pittsburgh, Pittsburgh, PA;Graduate Program of Information Science and Technology, University of Pittsburgh, Pittsburgh, PA

  • Venue:
  • SUM'11 Proceedings of the 5th international conference on Scalable uncertainty management
  • Year:
  • 2011

Quantified Score

Hi-index 0.00

Visualization

Abstract

Historical data reports on numerous events for overlapping time intervals, locations, and names. As a result, it may include severe data conflicts caused by database redundancy that prevent researchers from obtaining the correct answers to queries on an integrated historical database. In this paper, we propose a novel conflict-aware data fusion strategy for historical data sources. We evaluated our approach on a large-scale data warehouse that integrates historical data from approximately 50,000 reports on US epidemiological data for more than 100 years. We demonstrate that our approach significantly reduces data aggregation error in the integrated historical database.