gtg944s / STAT_COMP

Statistical Computing (STT 802, EPI853b)

STATISTICAL COMPUTING

This GitHub serves as a repository for the 2nd Half of the Statistical Computing Courses STT 802 and EPI-853b.

Instructors: Hyokyoung Grace Hong ([e-mail]) & Gustavo de los Campos ([e-mail])

Syllabus

Time & Place M/W 3:00 PM - 4:20 PM. Wells Hall B110F.

First Half [by Hyokyoung Grace Hong ([e-mail])]

Module 1: Introduction

1. R Studio/R Markdown
2. Introduction to R
3. Data preparation and descriptive analyses

Module 2: Statistical Models

1. Linear regression models (point estimates, SEs, t-test, p-values)
2. Generalized linear models (GLM)
3. Survival models
4. Quantile regression models

Module 3: High-dimensional Data Analysis

1. Regularized regression methods
2. Variable screening methods

Second Half [by Gustavo de los campos ([e-mail])]

Module 4: Sampling Random Variables

The following handout covers the topics that will be discussed in class

1. Pseudo random numbers
2. Transformation of random variables (see also a map of univariate distributions).
3. Inverse-probability method
4. Composition sampling
5. Gibbs sampler
6. Generating samples from the multivariate normal distribution

• In-class assgiment 1
• In-class assgiment 2
• HW3

Module 5: Estimating Power & Error Rate using Monte Carlo Methods

• Power and Type-I Error Rate
• In-class assigment 3
• Multiple Testing (& Family-wise error rate)
• In-class assigment 4

Moudle 6:Large-scale hypothesis testing

• Controlling type-I (family-wise) error rate
• Controlling false discovery rate (FDR)
• Inclass-5

Module 7: Resampling methods

• Permutation.
• INCLASS 6
• Bootstrap
• INCLASS 7 , INCLASS 8
• Cross-validation
• INCLASS 9

Module 8: Maximum likelihood via the EM-algorithm

• Introduction
• Example 1: right censored data
• Example 2: mixture models
• INCLASS 10

Bonus: Complex trait prediction using independent screening, Lasso, Elastic Net, and Bayesian models.

• Example 1
• Example 2