Errata / Corrigendum for Efficient Discovery of Approximate Dependencies Sebastian Kruse and Felix Naumann
PVLDB 11 (7), 759-772
Readers of the paper have pointed out a few minor errors, which we document here to ease the understanding
of the algorithm.
Erratum 1) In Section 5.1, the PLI for "Last name" should read 1,4, 3,5.
Erratum 2) In Section 5.3, Example 4, the tuple pairs (t1, t3), (t1, t5), and (t2, t3) should yield the
agree set sample AS = (, 1), (First_name, Town, 1), (ZIP, 1)).
Erratum 3) In Section 5.3, the example AUCC error of the attribute combination A1...An should be 0.0099.