Joeri Rammelaere

Member of Adrem Data Lab at the University of Antwerp
Office G.322
Middelheimlaan 1, 2020 Antwerpen, Belgium
E-mail: joeri.rammelaere(at)uantwerpen.be



Research


During a typical working day, I study problems related to rule-based data cleaning, with an emphasis on using pattern mining techniques for the discovery of said cleaning rules.



Publications

  • Cleaning Data with Forbidden Itemsets (ICDE 2017, Full Version, Poster)
  • Explaining Repaired Data with CFDs (VLDB 2018, Full Version, Presentation, Poster)
  • Revisiting Conditional Functional Dependency Discovery: Splitting the ā€œCā€ from the ā€œFDā€. (ECML/PKDD 2018, Full Version, Presentation, Poster)


    • Implementations

      • CFD Discovery code (paper "Revisiting CFD Discovery"):
        GitHub   CodeOcean
      • XPlode code (paper "Explaining Data Repairs With CFDs"):
        GitHub   CodeOcean
      • FBIClean code (paper "Cleaning Data with Forbidden Itemsets"):
        GitHub   CodeOcean
      • CFDMiner code (paper "Discovering Conditional Functional Dependencies" by Fan, Geerts, Li & Xiong):
        GitHub



      Experiment with Forbidden Itemsets