2. – 3.3. Workshop: Advanced Topics in Mixed Modeling (Dr. Dale Barr)

On Monday, 2nd of March and Tuesday, 3rd of March 2015, we will have a workshop.


Topic: Advanced Topics in Mixed Modeling (Dr. Dale Barr)

In this two-day workshop we will cover advanced topics in using mixed-effects modeling, including:

– specifying random effects
– coding categorical predictors
– interpreting continuous-by-categorical interactions
– generalized linear models for categorical/count data
– model checking and convergence problems

Participants are encouraged to bring in their own datasets for consideration.

Dr. Dale Barr is Senior Lecturer at the Interdisciplinary Study of Social Interactions and the School of Psychology at the University of Glasgow and is an expert on the psychology of communication and conversation, and statistical methodology. 

Schedule:
Monday, 2nd March: 9:00 a.m. – 5:00 p.m. (workshop)
Monday, 2nd March: 7:00 p.m. (informal get-together)
Tuesday, 3rd March: 9:00 a.m. – 3:00 p.m. (workshop)

The workshop takes place in Room 0.01 in Cartesium (Uni Bremen).

There are still places in this workshop. To register, please send an e-mail to vivien.mast@uni-bremen.de

Posted in Uncategorized

3.12. Presentation: Parallel Processing with R (Arne Hendrik Schulz)

On Wednesday, 3rd of December 2014, we will have a presentation. The meeting takes place at 10 a.m. (sharp/s.t.) in Room 0.41 in Cartesium (Uni Bremen).


Topic: Parallel Processing with R (Arne Hendrik Schulz)

Modern computers have processors with two or more cores. Most R functions, also computationally extensive ones, are single threaded, which means that they only use one core. There are several packages available that offer parallelization of processes to tackle this problem and to speed up computations. We will focus on the (do)SNOW package which works quite well with all operation systems (Windows, Mac, Linux) and can be used along with plyr. Applications are not limited to local workstations/laptops, they can also be executed on high-performance computing (HPC, supercomputing) clusters like the North-German Supercomputing Alliance (HLRN).

Posted in Uncategorized

19.11. Presentation: Object Oriented Programming with R (Mark Heckmann)

On Wednesday, 19th of November 2014, we will have a presentation. The meeting takes place at 10 a.m. (sharp/s.t.) in Room 0.41 in Cartesium (Uni Bremen).


Topic: Object Oriented Programming with R (Mark Heckmann)

R is a scripting language with object oriented features. You can see object orientation in action when you e.g. apply the function ‘plot’ to a vector or a regression object. The output is totally different. In the workshop we will look at what happens behind the scene and learn how to build S3 classes and methods.

Posted in Uncategorized

5.11. Paper Discussion: The Insignificance of Statistical Significance Testing

On Wednesday, 5th of November 2014, we will have a discussion meeting. The meeting takes place at 10 a.m. (sharp/s.t.) in Room 0.41 in Cartesium (Uni Bremen).

Topic: The Insignificance of Statistical Significance Testing (Douglas H. Johnson)

Now that we know so much about testing statistical significance, a text that tells us we’re doing it all wrong is just what we need! This is an invitation to reflect a little on what we’re doing in our day-to-day work, and what good science looks like. The text is in our dropbox folder.
If you don’t have access, let me know and I’ll send you the paper!

Posted in Uncategorized

4.6. Working Meeting: Data Mining in Social Media

On Wednesday, 4th of June 2014, we will have a working meeting. The meeting takes place at 10 a.m. (sharp/s.t.) in Room 0.41 in Cartesium (Uni Bremen).

Topic: Data Mining in Social Media

In this meeting, Niels will report his progress on accessing twitter data and analyzing it with R. Further, anyone is invited to present their own data mining efforts in social media, or bring ideas they want to work on. If possible, any data/scripts should be added to the group dropbox before the meeting. If you are not a member of the group yet, contact vivien.mast@uni-bremen.de for getting access to the dropbox folder.

Posted in Meetings, Working Meetings

21.5. Working Meeting: Mixed Models in Action

On Wednesday, 21st of May 2014, we will have a working meeting. The meeting takes place at 10 a.m. (sharp/s.t.) in Room 0.41 in Cartesium (Uni Bremen).

Topic: Mixed Models in Action

In this meeting, Emma will present the mixed models she fitted for her experimental data for discussion. Further, anyone is invited to present their own modelling efforts, or bring their data for us to work on together. Ideally, any data should be added to the group dropbox before the meeting. If you are not a member of the group yet, contact vivien.mast@uni-bremen.de for getting access to the dropbox folder.

Posted in Meetings, Termine, Working Meetings

9./10.4. Workshop: Grundlegende und fortgeschrittene Regressionsmethoden in R (Sascha Wolfer)

Unser Doktorandennetzwerk bietet am 9. und 10. April folgenden Workshop an, in dem noch einige Plätze frei sind:

Grundlegende und fortgeschrittene Regressionsmethoden in R (Sascha Wolfer)

Im zweitägigen Intensivworkshop “Grundlegende und fortgeschrittene Regressionsmethoden in R” werden die TeilnehmerInnen die theoretischen Grundlagen von Regressionsmethoden und deren einfache Anwendung in der statistischen Programmiersprache R erlernen. Weiterhin werden fortgeschrittene Methoden behandelt — insbesondere lineare gemischte Regressionsmodelle (linear mixed-effects models, hierarchical models). Je nach Interesse und Geschwindigkeit im Kurs wird außerdem auf logistische Regressionsmodelle und/oder Modelle mit nicht-linearen Prädiktoren eingegangen.

Der Schwerpunkt wird auf der Anwendung im wissenschaftlichen Alltag liegen. Es ist Zeit für Nachfragen, Diskussionen und die Visualisierung von Modellverhersagen eingeplant. Voraussetzungen sind das Wissen über statistische Basiskonzepte (Maße der zentralen Tendenz, Dispersionsmaße, Korrelation) und zumindest grundlegende Kenntnisse im Umgang mit R. Jede/r Teilnehmer/in sollte im Kurs Zugang zu einem Computer mit R (www.r-project.org) haben. Es können gern eigene Datensätze in den Kurs mitgebracht werden, anhand derer einige der Konzepte erläutert werden könnten.

Zum Dozenten: Sascha Wolfer ist Mitarbeiter im Projekt Empirische Methoden der Abteilung Lexik am Institut für Deutsche Sprache (IDS), Mannheim und führt regelmäßig Seminare zu statistischen Methoden mit R durch.

Termin: 9. und 10.4., 9:00 – 17:00 Uhr, Raum: Cartesium, 0.01

Anmeldung: verbindliche Anmeldung per e-mail an Vivien Mast (vivien.mast@uni-bremen.de), mit Betreff “registration” und Angabe des vollständigen Namens in der e-mail. Es gilt das Prinzip first come – first serve.

Posted in Termine, Workshops