STAT 394A - Statistical Learning Workshop

4 Credit(s)

This is an applied workshop in which students will learn advanced statistical techniques to analyze datasets. Topics will include non-linear regression, classification, regularization, smoothing, model selection, model assessment, model inference, bootstrapping, neural networks and flexible discriminant analysis. Emphasis will be placed on developing the mathematical underpinnings of each technique, writing code in MATLAB for numerical implementation and interpreting the results in the context of scientific hypothesis testing. STAT 217 Intermediate Statistics is a pre-requisite for this class so entering students will be familiar with data collection, analysis, and interpretation of results in the framework of scientific hypothesis testing.

This is a project-based class in which each student will propose a project, design data collection algorithms to gather data from a large database (such as the human genome), analyze the data using appropriate statistical learning techniques, present project updates, give an oral presentation of final results and write a formal paper. In most cases, this course will be themed so that real data sets are taken from a specific application area (such as ecology, genetics, geology, climatology, etc.) and are analyzed and/or modeled using appropriate techniques. In light of this, the specific techniques students use will vary from class to class

Prerequisite(s): M 210 and STAT 217 or STAT 233 grade B- or higher; or c/i.
(fall/odd-numbered year).