Webcasts

Robust predictive modeling through Structural Risk Minimization
Date: May 15, 2007
Time: 9:00 am PST/11:00 am CST/12:00 noon EST

Guest Speaker: Dr. Isabelle Guyon, expert in machine learning theories, data mining, statistical data analysis, and predictive modeling, and one of the three holders of the SVM (Support Vector Machines) US Patent

Why Structural Risk Minimization (SRM)?

The primary challenges for statisticians are to build highly accurate models that are also reliable. This has been particularly difficult when confronted with large number of variables as it is often the case nowadays. The more variables there are, the more difficult it is to build reliable models. The more variables, the more time and expertise are needed.

Also, predictive models should not just fit the training examples, but they must also make good predictions on new data. Good learning algorithms avoid "overlearning" the training data or "overfitting".

Under these circumstances, how and why SRM can help?

SRM, developed by Vladimir Vapnik, is a breakthrough in mathematics and statistics that for the first time makes it possible to automatically build reliable and accurate models in high dimension space. In contrast to traditional statistics models, SRM models become more accurate and are still reliable as the number of variables is increased. Model Accuracy and Reliability are determined by the data. This becomes of great help even for expert users.

Join us and our special guest speaker, Isabelle Guyon, one of the pioneers of the use of SRM, for this live 60 minute web seminar explaining the Structural Risk Minimization (SRM) principle and how it is put to work to avoid the problem of "overfitting".

Register Now!

What's this?