Doctoral Researcher at the Autism, Social Cognition & Bodily Movement Lab (jencooklab.com), Senior Centre Technician at @TheCHBH, MATLAB Ambassador, He/Him.
š New paper alert! š§ The velocity-curvature power law is one of the 'kinematic laws of nature' (Flash 2021), a 'fundamental law of human control' (Zago et al. 2016), and one of the 'kinematic regularities' (Frith & Frith 2023) thought to underlie all biological motion. 1/9
Follow-up to my ICC thread from last week š§µ I wanted to test Salarian's MATLAB FileExchange code. Thanks to @cchapman.bsky.social for pointing me toward R. I took it further than perhaps intended.
New experience today š: I set up an experiment to check the testāretest stability of my favourite kinematic measure: Lacquanitiās 1/3 power law. A velocityācurvature relationship that results in a kinematic invariance which might underlie all biological motion.
You can even add the necessary packages without ever looking at R
system('Rscript -e "install.packages(''psych'', repos=''cran.r-project.org'')"')
system('Rscript -e "install.packages(''ICC'', repos=''cran.r-project.org'')"')
#MATLAB#MATLABambassador#R
New experience today š: I set up an experiment to check the testāretest stability of my favourite kinematic measure: Lacquanitiās 1/3 power law. A velocityācurvature relationship that results in a kinematic invariance which might underlie all biological motion.
But⦠last updated in 2016. How far do we trust ārandomā code? š¤ I will havw to write some tests. This is the reality of (apparently!) niche stats in MATLAB, community-driven, not MathWorks-authored. Who do I have to bribe to get it added?
So hereās the question: whatās your ICC workflow? There must be dozens of us! Dozens!
MATLAB, R, Python⦠or do you still cling to SPSS? š§
#MATLAB#MATLABambassador
š§µ Back in 1999, I was hand-rolling the Karhunen-LoĆØve transform using eig() ⦠you may know it as Principal Component Analysis, or as the cool kids now know it by its rapper name PCA. š¤
I had no fancy toolboxes with the pca() command.
š It lets you: ⢠Manipulate μ and Ļ of up to 3 variables ⢠Adjust correlations ⢠Watch Principal Components (PCs) emerge in real-time ā”You can rotate 3D plots š, click individual points to see their deviation vectors animated, and visualize projections onto each PC.
Available on MATLAB Onlineā just click that Launch button on GitHub. š You donāt even need a license - 20hrs/month free!
#MATLAB#PCA#KarhunenLoeve#MATLABambassador