[Pols-announce] Fwd: [Crmda-l] Workgroups Forming for Fall 2018

[Paul Johnson]

We are starting a new workgroup, which I'm calling "Too many predictors" 
behind its back.  It is under leadership of Business professor Ben 
Sherwood.  This will touch on things that are truly of interest to many 
social scientists, such as

- what's wrong with stepwise regression
- can regression "regularization" give us a get of jail free card 
(ridge, lasso, elasticnet)?
- will AIC and BIC "information" criteria offer meaningful guidance

I hope you will consider coming for our planning session for the 
workgroups on August 24th.


-------- Forwarded Message --------
Subject: 	[Crmda-l] Workgroups Forming for Fall 2018
Resent-Date: 	Thu, 16 Aug 2018 13:38:26 -0700 (PDT)
Resent-From: 	pauljohn at ku.edu
Date: 	Thu, 16 Aug 2018 20:38:21 +0000
From: 	CRMDA-L <crmda-l at lists.ku.edu>
Reply-To: 	CRMDA. <admin-crmda at ku.edu>, CRMDA-L <crmda-l at lists.ku.edu>
To: 	CRMDA-L <crmda-l at lists.ku.edu>, METHODS-L <methods-l at lists.ku.edu>, 
KUANT-L <kuant-l at lists.ku.edu>



Hello All,

I hope everyone enjoyed their summer and has a good first week back next 
week! As always, CRMDA will be hosting Workgroups this Fall. We have 
created a new group this year as well. Below are the details or you can 
visit the Workgroup website: *http://crmda.ku.edu/workgroups*.

To help us determine the best schedule and topics for these Workgroups, 
we will be hosting an organizational meeting next *Friday, August 24th 
from 2-3 p.m. in Watson Library room 455*. Anyone is free to participate 
in this discussion. Please bring ideas with you!

*_Below are the descriptions for each Workgroup:_*

·*Bayesian Multilevel Modeling Workgroup – Paul Johnson, CRMDA Director*

oThis group will develop examples and workshop notes for an overview of 
Bayesian research methods and applications to multi-level ("mixed 
effects") models.**

·*Big Data: Analysis with Many Predictors – Ben Sherwood, CRMDA Faculty 
Fellow, Assistant Professor, School of Business*

oA common problem in the big data era is how to deal with a large number 
of predictors. In some cases the number of predictors can be larger than 
the sample size, thus making it impossible to fit classical models, such 
as least squares, without doing some variable selection. Topics 
discussed in this Workshop will include: Variable screening, methods for 
balancing fit and model size (AIC, BIC, cross-validation) and penalized 
regression**

If you have any questions, please let me know.

Thank you!

*Auburn Packer*

Administrative Associate

Center for Research Methods & Data Analysis

College of Liberal Arts & Sciences | University of Kansas

Watson Library, 470B

1425 Jayhawk Blvd | Lawrence, KS  66045

785.864.3353 | crmda.ku.edu | crmda at ku.edu <mailto:crmda at ku.edu>


-- 
Paul E. Johnson			University of Kansas	     	
Professor			Director, Center for Research
Political Science		Methods & Data Analysis (CRMDA)
http://pj.freefaculty.org	http://crmda.ku.edu
email: pauljohn at ku.edu
Address: CRMDA
	 Watson Library, Suite 470
	 1425 Jayhawk Blvd.	
    	 Lawrence, Kansas
  	 66045-7594
   	 Ph: (785) 864-8215

-------------- next part --------------
_______________________________________________
Crmda-l mailing list
Crmda-l at lists.ku.edu
https://lists.ku.edu/listinfo/crmda-l