15th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining (KDD), Paris 2009
The annual ACM SIGKDD conference is the premier international forum for data mining researchers and practitioners from academia, industry, and government to share their ideas, research results and experiences. KDD-09 will feature keynote presentations, oral paper presentations, poster sessions, workshops, tutorials, panels, exhibits, demonstrations, and the KDD Cup competition.
External links:
Invited talks | ||||
ACM SIGKDD Award Presentations | ||||
SIGKDD Annual Business Meeting | ||||
I enjoyed all conferences, however specifically I would like to ask to Usama Fayyad and to Dr Grossman, wht do you thoink about to use the error of varianza instead of the model variance to approach the question problem because of the error variance is more consistent and I am not sure that this part needs to many adjustment that usually we use to turn countinuous nonrectilinear variables in @forced rectininear voariables to be submmited in a linear regression model. In a multivariable analsis we fit de model variance depending on the number of variables using different kind of conditional *Akai, Kendall etc* why not make data mining of residual instead of modelling test_-_