Stat Learning & Data Mining

Hours

3.0 Credit, 3 Lecture, 0 Lab

Prerequisites

Semester

Fall

Multiple linear regression, nonlinear regression, local regression, penalized regression, generalized additive models, logistic regression, discriminant analysis, tree-structured regression, support vector machines, neural networks.

STAT 536

This course trains students in using statistical methods for modeling a response variable as a function of explanatory variables. Stat 535 covers linear models, and this course attempts to cover the complement set. At a minimum you will learn the derivation, computation, and application of the different methods on data.

Linear Regression

Review Linear Regression Models

Weighted Least Squares

Review Weighted Least Squares, Mixed Models

Bayesian

Bayesian Linear Regression

Measurement

Measurement Error Models

Linear Models

Generalized Linear Models (logistic)

Model Assessment

Model Assessment and Selection

Shrinkage Methods

Shrinkage Methods, Bias-Variance Tradeoff, Subset Selection

Local Regression

Local Regression (splines, smoothers)

GAM

Generalized Additive Models (GAM)

Tree-based Models

Tree-based Models, Random Forests

Boosting

Boosting, Bayesian Adaptive Regression Trees