Short Courses

See here for the Friday Short Courses Schedule.

Short Course 1

Instructor: Cyrus Mehta and Charles Liu,
Affiliation: Cytel, Inc
Title of Speech: Adaptive Design for Confirmatory Clinical Trials
Location: FXB-G13
Time: 9:00am-5:30pm
Lunch Break: 12:30pm-2:00:pm

General Course Summary:

This workshop will focus on strategies for efficient decision making via interim analyses of ongoing clinical trials using state of the art methods for group sequential and adaptive designs. Topics covered will include group sequential design and monitoring with sample size re estimation, preserving type 1 error, computing power, obtaining point estimates and computing confidence intervals in the adaptive setting. Case studies in oncology and cardiology are used to reinforce the main points. The workshop includes a hands on session with the EAST® 6.2 software.

Short Course 2

Instructor: Han Liu
Affiliation: Princeton University
Title of Speech: Analysis of Big Data
Location: FXB-G12
Time: 9:00am-5:30pm
Lunch Break: 12:30pm-2:00:pm

General Course Summary:

The amount of data in our world is exploding and analyzing large datasets is becoming central in science, technology, and our society. This course introduces new statistical methods and computational tools for analyzing Big Data. This course covers the following topics:
1. Statistical challenges of Big Data Analysis
2. Handling massive data: divide and conquer statistical inference
3. Handling high dimensional data: regularization method
4. Handling complex and noisy data: robust semiparametric method
5. Large scale statistical optimization: model based optimization

The main focus of this course is to introduce new methodology and explain theoretical justification of these methods at an intuitive level.

For a syllabus of this course, please click here for Big Data Syllabus .

Click here forBig Data Part 1Big Data Part 2Big Data Part 3Big Data Part 4

Short Course 3

Instructor: John Buonaccorsi, Professor Emeritus
Affiliation: University of Massachusetts‐Amherst
Title of Speech: Measurement Error and Misclassifi cation
Location: Kresge 212
Time: 9:00am-5:30pm
Lunch Break: 12:30pm-2:00:pm

General Course Summary:

Measurement error is ubiquitous and it is well known that the inability to exactly measure either outcomes or predictors often leads to biased estimators and invalid inferences. This problem has received considerable attention of late across many areas, especially in epidemiology. This course will present an introductory and relatively applied look at measurement error with a focus on models used for measurement error, the impacts on naive analyses that ignore it, and an overview of techniques to correct for it. Emphasis is on misclassifi cation in basic categorical models (estimating proportions and in two-way tables) and measurement error in predictors in linear and non-linear regression (including logistic regression). In the regression contexts we describe a variety of correction methods that are used with either replication or validation data. A number of illustrative examples from various disciplines will be presented and an overview of software options will be provided. While we present a little of the basic theory the course focus is primarily applied so it is accessible to a broad audience and allows us to cover more topics in a limited amount of time. Students should have at least some prior exposure to basic categorical methods (single proportions and two way tables) and linear and logistic regression and be familiar with seeing models and methods expressed in matrix-vector form.

No book is required although the lecture notes come mainly from Buonaccorsi (2010), Measurement Error: Models, Methods and Applications”, Chapman & Hall.