Fabian M. Suchanek Rule Mining 140
© TV week © Dailymail © Public Domain CC by Jean-Marc Ayrault CC by Matthieu Riegler © Voici Rules hasChild citizenOf hasChild marriedTo marriedTo citizenOf presidentOf hasChild hasChild hasChild hasChild 2
© TV week © Dailymail © Public Domain CC by Jean-Marc Ayrault CC by Matthieu Riegler © Voici Rules 3 hasChild citizenOf hasChild marriedTo marriedTo citizenOf presidentOf hasChild hasChild hasChild hasChild “Whenever someone has a child, this is also the child of the spouse” marriedTo hasChild hasChild
© TV week © Dailymail © Public Domain (rules don’t have to be always correct) Rule Mining hasChild citizenOf hasChild marriedTo marriedTo citizenOf presidentOf hasChild hasChild 4
Rule Mining 5 hasChild citizenOf hasChild marriedTo marriedTo citizenOf presidentOf hasChild hasChild ? >uses © TV week © Dailymail © Public Domain
Rule Mining for Completion 6 marriedTo hasChild >uses
Rule Mining for Completion 7 marriedTo hasChild hasChild >uses
Rule Mining for Correction 8 marriedTo hasChild hasChild >uses
Rule Mining for Correction 9 marriedTo hasChild hasChild >uses
... obvious for us, but non-trivial for a computer. Applications in •   artificial intelligence / question answering •   information extraction (see  lecture ) •   science •   engineering (fault finding) •   medicine •   social network mining Rule Mining for Insights 10 ... in 4% of the cases