Requests

Final completion of the Ph.D. is almost in sight.  I am on the search
for some data with a time stamp attribute that has got concept drift
occurring within it to complete the experimentations.

The on-line learning algorithm being used is a tree induction
classifier which detects concept drift in the presence of noise and
then removes invalid examples.  A large database would be preferable
to aid extraction of batches.

Reference of work to date:

Black, M., Hickey, R.J. "Maintaining the Performance of a Learned
Classifier under Concept Drift", Intelligent Data Analysis 3 (1999)
453-474.

Hickey, R.J., Black, M., "Refined Time Stamps for Concept Drift
Detection", Spatio-Temporal Data Mining - TSDM2000, published in LNAI
2007 20-30, Springer-Verlag.

Please reply to mm.black@ulst.ac.uk