Description:

We will continue from the previous lectures and mainly focus on two topics: 1. Demonstrations of model selection and model averaging methods with R; 2. Model assessment from goodness of fit perspectives. For the latter, when models are used for classification, it is important to note that the currently dominating practice of using classification accuracy to evaluate a model may not be proper in terms of addressing the question whether the model can be improved or not regardless of how accurate (or inaccurate) it may be. We will present recent developments in this area.

Lecturer: 

Yuhong Yang, Professor, School of Statistics, University of Minnesota