Skip to content

Latest commit

 

History

History
38 lines (27 loc) · 3.4 KB

README.md

File metadata and controls

38 lines (27 loc) · 3.4 KB
Raw

posterSML

SML = Supervised Machine Learning.

The poster (see SMLPrinciples.pdf) was presented at the International Convention of Psychological Science (ICPS) 9th - 11th March 2023. The poster was presented on Thursday, 9th March (11:45 - 12:45 h) at Panoramic Hall (Location: SQUARE - Brussels Convention Centre).

The poster is based on this manuscript which is currently under review:

Miché, M., Zander-Schellenberg, T., Wahl, K., & Lieb, R. (2022). A demonstration of supervised machine learning for psychological researchers [Manuscript submitted for publication]

What do you want to know next about SML?

I provide diverse SML input and/or files which you might be interested in (see main posterSML folder). If you have a particular request or question that has not yet been covered, use the 'Issues' tag above. Please note: I provide this 'service' based on my intrinsic motivation to be of help (and to be helped). Anybody who decides to use this GitHub page for issuing ideologic preferences, e.g., use this instead of that software, will not receive any attention (or answer) from me, because to me that is a waste of time and energy.

Q&A

A collection of questions which in hindsight I would have liked to have an answer to when I was an SML newbie. As with any answer, the answers I provide reflect my current point of view. Other people may see things differently.

SML recommended reading

Just import the SMLrecommendations.bib document into your reference management software, e.g., zotero. This is of course only a very short list of articles. Browsing the web for 'supervised machine learning' (introduction or tutorial) will provide plenty of information to select from.

Hands-on SML

You can start experimenting with SML right away by downloading and using the supplementary material of the manuscript which this poster is based upon.

If you know how to use R, you can download the supplementary material within R like this:

# Only required if you have not installed the devtools package yet.
install.packages("devtools")
# Install demoSML from GitHub
devtools::install_github(repo="https://github.com/mmiche/demoSML",
                      dependencies = "Imports", build_vignettes = TRUE)
# Load the demoSML package
library(demoSML)
# Open the package documentation, notice the link 'User guides, package vignettes ...'.
help(package="demoSML")

Which software to use for SML

Use whatever software you are most comfortable with. I use R, like (probably) most psychologists nowadays. The advantage of R may at the same time be its disadvantage: It is free. Everybody can contribute to it, meaning that everybody can write a software package, upload it to CRAN and/or to GitHub, from where everybody can download and then use it. Therefore, several individuals and teams provide 'their' SML software solutions within R. For instance, currently two competitors are Tidymodels and mlr3. Both provide an interface to a very large number of SML relevant R software packages. Due to each single package having idiosyncrasies of its developer(s), e.g., column.names, column_names, colNames, ..., using an interface may be a good decision in the long run, because of its consistent use of commands.