Optimal rule discovery and applications
Speaker: Jiuyong Li, USQ
When: 2006-08-24 12:00:00
Venue: D109, Unviersity of Southern Queensland, Towoomba
Host: Michael de Raadt
Abstract:Rules are one of the most expressive and human understandable
representations of knowledge; a rule based method produces
self-explanatory results. Therefore, rule discovery has been a major
issue in machine learning and data mining. Heuristic and association
rules are two dominant schemes for rule discovery. However,
heuristic rule discovery fails to find many globally optimal
rules. Association rule discovery yields too many rules and is
infeasible for many applications. In this talk, I will present an
efficient alternative for rule discovery, optimal rule discovery,
and discuss its relationships with other rule discovery schemes. I
will also discuss its applications. One application is to build
classification systems that tolerate missing values. Another
application is to find risk patterns in a real world medical data.
Biography:(biography unavailable)
Type: USQ seminar
Contact:Michael de Raadt, seminar host (deraadt@usq.edu.au)
or Guido Governatori (ITEE seminar co-ordinator)
(guido@itee.uq.edu.au)