7 Dec: Presentation @ Data Science Seminar in Leiden

Geplaatst 14 nov. 2018 14:02 door Marjolein Fokkema   [ 14 nov. 2018 14:04 bijgewerkt ]

I will be presenting my research at the Data Science Seminar at the Faculty of Social Sciences, Leiden University, December 7th 15.00-17.00. Hope to see you there!

pre 0.6: Prediction rule ensembles for survival outcomes

Geplaatst 21 aug. 2018 05:45 door Marjolein Fokkema

pre version 0.6 is now available from CRAN! Key improvements include support for survival responses and shorter computation times.

Let rules rule with pre 0.5!

Geplaatst 10 mei 2018 18:06 door Marjolein Fokkema

Version 0.5 of package pre (and R package for fitting prediction rule ensembles) is now available from CRAN. The new and improved version supports continuous, binary, multinomial, count and multivariate continuous response variables and allows for selecting from a wider range of rule learning procedures and parameters. So stop hanging out in the sunshine, get behind your computer and let rules rule!

June 24: Workshop on mixed-effects trees at UCL / TSIL2018

Geplaatst 8 mrt. 2018 07:46 door Marjolein Fokkema

On June 24, at the Treatment Selection Idea Lab (TSIL) 2018 at University College London, I will be teaching a pre-conference workshop on fitting mixed-effects regression and classification trees. I will show how these mixed-effects trees can be used for treatment selection and treatment outcome prediction in datasets with and without multilevel structures.

You can register for the conference and workshops here: https://onlinestore.ucl.ac.uk/conferences-and-events/faculty-of-brain-sciences-c07/clinical-educational-and-health-psychology-f66/f66-treatment-selection-idea-lab-tsil-conference-2018

glmertree paper accepted in Behavior Research Methods

Geplaatst 5 sep. 2017 08:55 door Marjolein Fokkema

Our paper on glmertree (an R package for fitting trees on clustered/multilevel datasets) and treatment-subgroup interactions has been accepted for publication in BRM! I am grateful for the helpful reviewer comments, the paper and package have been much improved since the 2015 working paper. Click here to read the pre-print!

New version of pre (0.4) available on CRAN

Geplaatst 5 aug. 2017 04:30 door Marjolein Fokkema

The latest version of pre (an R package for fitting prediction rule ensembles) is now available on CRAN. It's faster, many bugs have been fixed and functionality has been added. For more info, check out the examples on github or the working paper on arXiv.

International Short Visit granted by the Swiss National Science Foundation

Geplaatst 6 mei 2017 02:44 door Marjolein Fokkema

I was very happy to receive news today that the SNSF awarded me an International Short Visit grant to support my visit to Prof. Strobl's lab at the University of Zurich. I'm grateful for SNSF's confidence in our research on recursive partitioning methods!

May 15-16: Course on trees @ Zurich R courses

Geplaatst 11 apr. 2017 04:49 door Marjolein Fokkema

I'll be teaching a course on recursive partitioning methods at the Zurich R courses (a continuing education program of UZH and ETH Zurich).

For more info on the course (titled 'Introduction to classification and regression trees, random forests and model-based recursive partitioning in R'), see: http://www.zhrcourses.uzh.ch/en/programm/intro-trees.html

April 27: Talk @ ZüKoSt (Seminar on Applied Statistics)

Geplaatst 11 apr. 2017 04:43 door Marjolein Fokkema

I'll be giving a talk on prediction rule ensembles and my R package 'pre' at ETH Zurich. For more info, see:
https://www.math.ethz.ch/sfs/news-and-events/seminar-applied-statistics.html

Visiting researcher at UZH

Geplaatst 27 feb. 2017 06:03 door Marjolein Fokkema

From March until August 2017, I am a visiting researcher at the lab of Prof. Dr. Carolin Strobl at the University of Zurich

A joyous occasion indeed, as I will collaborate with the leading group in tree-based methods and psychometrics, while enjoying beautiful views of the Zürichsee and the Swiss Alps.

1-10 of 14