Knowledge Discovery and Data Mining (KDD) 2012 Beijing Recap
It was great to be in Beijing for the 2012 KDD meetings where we managed to meet up with old friends and meet new like-minded people during three well-organized long days. Salford was a sponsor and we had the opportunity to demonstrate both the English and the Chinese-language versions of the SPM Salford Predictive Modeler software suite at our exhibitor booth.
The meetings included formal papers, panel discussions, and exchanges which were very useful for taking the pulse of the industry and the discipline of "knowledge discovery from data." Our plan is to offer some summary and discussion of a few of the papers and sessions we think would be of greatest interest to our (present and future) user community. We certainly learned a great deal at the meetings but we were also pleased to observe that several ideas that we have been developing over the last decade or more are being rediscovered by other researchers.
At Salford we are accustomed to being at times several years ahead of the pack because of our active and ongoing discussions with Jerry Friedman (and during his lifetime, Leo Breiman). Also because we have been wrestling with tough data mining problems for over 20 years and have managed to come up with some promising solutions.
Topics that we plan to discuss in the near future:
"Big Data" including some observations on the panel discussion on big data and papers dedicated to the topic; using TreeNet to improve logistic and multiple regression models; intelligent ad serving in an ad network;cost-sensitive learning; and, some comments on the KDD Cup 2012 where one of our clients did very well using TreeNet models.
We are glad you are visiting our web site and blog and hope you continue to find the posts valuable.
Watch our Big Data Video Series!