Beyond P-values 2017 Regression Analysis

 

beyond p bear

Image: Courtesy of Philippe Naveau

The Beyond P-values course series is aimed at NCAR  staff and visitors with an interest in acquiring more advanced statistical and data analysis skills. Each course will provide a targeted selection of topics in a  specific statistical subject taught by experts in that area. The instruction is data-driven and includes hands-on teaching of statistical software tools. The courses last two to three days and are limited to a small number of participants to provide high quality one-on-one support by the instructors and additional coaches. For questions and suggestions of topics, please contact Dorit Hammerling (dorith@ucar.edu).

 

March 7 - 9, 2017 at NCAR

Course Description

This three-day course, March 7-9, 2017, will provide an introduction to more advanced topics within the area of regression analysis. After a review of regression basics, we will cover diagnostic tools and a variety of variable selection methods including stepwise selection and Lasso. Then we will introduce quantile regression, which aims at estimating the median or other quantiles, compared to traditional regression targeting the mean estimate. The course is aimed at an applied audience and will make heavy use of data examples to illustrate the concepts. We'll use the open-source statistical software R [https://www.r-project.org/]. The format of the course is hands-on and participants will use their own laptops.

Instructors

The lead instructor for the course is Brian Reich, Professor of Statistics at North Carolina State University. He will be assisted by Halley Brantley (North Carolina State University), Matt Edwards (Newcastle University), and Mark Risser (Lawrence Berkeley National Laboratory). Seats are limited to 12 participants to allow for effective one-on-one coaching. To apply, please visit the Regression Analysis Application link on the left hand-side of the page.

Tuesday, March 7, 2017, NCAR Mesa Lab - Damon Room (239)

 

 9:00  – 10:15 

 Review of linear regression

 10:15  – 10:45 

 Break

 10:45  – 12:00 

 Important cases of multiple linear regression

 12:00  – 1:00 

 Lunch

 1:00  – 2:15 

 Regression diagnostics

 2:15  – 2:45 

 Break

 2:45  – 4:00 

 Regression for non-Gaussian data

 

 

Wednesday, March 8, 2017, NCAR ML- Damon Room (239)

 

 9:00  – 10:15 

 Dealing with collinearity

 10:15  – 10:45 

 Break

 10:45  – 12:00 

 Variable Selection

 12:00  – 1:00 

 Lunch

 1:00  – 2:15 

 Penalized regression

 2:15  – 2:45 

 Break

 2:45  – 4:00 

 Non-linear regression

 

Thursday, March 9, 2017, NCAR ML - Damon Room (239)

 

 9:00  – 10:15 

 Dealing with spatiotemporal correlation

 10:15  – 10:45 

 Break

 10:45  – 12:00 

 Quantile regression

 12:00  –  1:00 

 Lunch

 1:00  – 2:15 

 Further topics in quantile regression

 2:15  –  2:45 

 Break

 2:45  – 4:00 

 Estimating extreme quantiles