The Database Group @ University of Toronto Midas Touch: Discovering Linkage Points over Web Data
     Project DescriptionPeople | Publications 
  New! Datasets and Experimental Results available here.
A web demo for online linkage point discovery will be made available soon.
   Project Description
   Discovering links within and between data sources is a challenging problem and attractive research area. Existence of links add value to data sources, enhance data and information discovery and allow or enhance many increasingly important data mining tasks. Without proper links, data sources resemble islands of data (or data silos), where each island maintains only part of the data necessary to satisfy a users' information needs. In such settings, apart from finding which data are of interest to them in a given island, users are also responsible for finding which other islands might hold relevant data, and what the specific pieces of relevant data are within these islands. The goal of the Midas Touch project is to assist users and data publishers in the very first step of their link discovery process: discovering linkage points, i.e., discovering the attributes that are shared or related, and can be used to effectively link the records between the data sets.
  • Discovering Linkage Points over Web Data
    O. Hassanzadeh, Ken Pu, S. Hassas Yeganeh, R.J. Miller, M. A. Hernandez, H. Ho, L. Popa
    Under submission
   Related Projects
  • Midas:  Intelligent Information Integration
  • xCurator:  Linking Semistructured Web Data
  • LinQuer:  Linkage Query Writer
  • Stringer:  Duplicate Detection System for String Data 
   Datasets and Experimental Results
Copyright © 2011 The Database Group @ University of Toronto | Last Updated November 28th, 2011