I was invited by my colleague Marcus Kubsch at IPN in Kiel, Germany to lead a workshop on machine learning in science education research as part of the Machine Learning and Computer-Based Text Analysis conference.
The workshop description is here:
Machine learning includes a wide range of methods and software tools, but it is possible to get started with applying machine learning methods in relatively short order through the use of R. In this workshop, researchers and analysts will learn about a) a small but important set of core ideas about machine learning, b) how these ideas are instantiated through packages (add-ons) for the R statistical software, especially the R packages parsnip, recipes, and tidymodels and c) how to conduct a start-to-end machine learning analysis in the context of science education research using the core ideas and R packages introduced. While some experience and prior use of R is required (an excellent introduction to R can be found here: https://r4ds.had.co.nz/index.html; one in the context of education is here: http://datascienceineducation.com/), participants are not required to have any prior knowledge or experience with machine learning methods.
Materials (the slides, especially) may “stand on their own” well and are here:
- Getting Ready (a primer on using R/the tidyverse; stands on its own pretty well) task: https://github.com/jrosen48/ML-in-Science-Education-Workshop-Materials/blob/master/getting-ready
- Slides: https://jrosen48.github.io/ML-in-Science-Education-Workshop-Materials/
- Machine Learning Regression task: https://github.com/jrosen48/ML-in-Science-Education-Workshop-Materials/blob/master/regression-ml.R
- Machine Learning Classification task: https://github.com/jrosen48/ML-in-Science-Education-Workshop-Materials/blob/master/classification-ml.R
A repository with the materials (and the data) are here: https://github.com/jrosen48/ML-in-Science-Education-Workshop-Materials