Correlated Component Regression - Revolutionising Predictive Modelling
Monday 3rd September
Holiday Inn, Telford, UK
Running on the eve of the Royal Statistical Society's 2012 Annual Conference, the course starts by exposing the problems and limitations of traditional linear and logistic regression when dealing with small sample sizes and/or near multicolinearity. Practical examples will demonstrate how CCR produces reliable predictions and improved interpretation, even when the number of predictors exceeds the number of cases. An overview of the relevant theory and comparison with other methods will be provided, and case studies with real and simulated data will be used to illustrate the method and its benefits. Exercises will be based on a demo version of CORExpress®, an extremely intuitive software programme that implements the theory of CCR into practical applications.
The course is aimed at statisticians, analysts and researchers who work with real data and want to discover how to add an accurate predictive modelling method to their statistical tool-kits.
Course delegates will discover how CCR can be used to:
Maximise Out-of-Sample Performance (ability to predict new cases)
Produce simpler, more interpretable models
Enable very robust models to be built with minimum demands on the data
Course Leaders:
Dr Jay Magidson, Founder of Statistical Innovations and pioneer of Latent Class Analysis
Gary Bennett, Founder of Logit Research
For more information and details of how to book see http://www.logitresearch.com/Training_and_Development
You may leave the list at any time by sending the command
SIGNOFF allstat
to [log in to unmask], leaving the subject line blank.
|