#CrossValidation
LightNews LightNews
kordinglab.bsky.social
Statistically crossvalidation folds are not independent. The number of papers treating them as if they were is off the charts.
bttyeo.bsky.social Thomas Yeo @bttyeo.bsky.social · May 21
In a meta-analysis of 210 biomedical AI studies that statistically compared models under cross-validation, 97% used invalid statistical tests.

Here's our new preprint doi.org/10.64898/202... led by @tianchu.bsky.social @hetuli.bsky.social @shaoshiz.bsky.social @nichols.bsky.social 1/N
May 21, 2026 at 1:11 PM
teeeeeeeee.bsky.social
`crossvalidation` on R-universe, plus a classification example
https://thierrymoudiki.github.io/blog/2021/07/23/r/crossvalidation-r-universe

#Python #DataScience #MachineLearning #rstats #Techtonique
September 25, 2026 at 7:43 PM
annalangener.bsky.social
✨ Excited to share our new preprint: osf.io/preprints/os... ✨

In this paper, Nick Jacobson and I dive into the challenges researchers face when developing prediction models for just-in-time adaptive interventions (JITAI). 🧵👇

#EMA #JITAI #CrossValidation #PassiveSensing
March 5, 2025 at 5:31 PM
kordinglab.bsky.social
Learning is for the future. So crossvalidation is a bit weird. open.substack.com/pub/kording/...
Prospective Learning: Anticipating a Changing Future
Trying to write down what learning is about in the real world.
open.substack.com
March 11, 2025 at 5:59 PM
teeeeeeeee.bsky.social
Time series cross-validation using `crossvalidation` (Part 2)

https://thierrymoudiki.github.io/blog/2021/11/07/r/misc/timeseries/crossvalidation-eval-ts

#Techtonique #DataScience #Python #rstats #MachineLearning
Blog post image
February 25, 2025 at 9:06 AM
teeeeeeeee.bsky.social
Time series cross-validation using `crossvalidation` (Part 2)

https://thierrymoudiki.github.io/blog/2021/11/07/r/misc/timeseries/crossvalidation-eval-ts

#Techtonique #DataScience #Python #rstats #MachineLearning
Blog post image
September 15, 2025 at 8:29 PM
teeeeeeeee.bsky.social
Time series cross-validation using `crossvalidation` (Part 2)

https://thierrymoudiki.github.io/blog/2021/11/07/r/misc/timeseries/crossvalidation-eval-ts

#Techtonique #DataScience #Python #rstats #MachineLearning
Blog post image
March 1, 2025 at 2:42 PM
teeeeeeeee.bsky.social
`crossvalidation` and random search for calibrating support vector machines
https://thierrymoudiki.github.io/blog/2021/08/06/r/crossvalidation-svm-r

#Python #DataScience #MachineLearning #rstats #Techtonique
July 16, 2026 at 11:03 PM
kroyindia.bsky.social
On a strategy of selecting the best machine-learning-based regression models from several competing models using cross-validation and (internal) validation set performance
doi.org/10.26434/che...
#MachineLearning #QSAR #Validation #CrossValidation #ExternalSet #Prediction #Metrics #Cheminformatics
June 16, 2026 at 8:34 AM
teeeeeeeee.bsky.social
Time series cross-validation using `crossvalidation` (Part 2)

https://thierrymoudiki.github.io/blog/2021/11/07/r/misc/timeseries/crossvalidation-eval-ts

#Techtonique #DataScience #Python #rstats #MachineLearning
Blog post image
November 30, 2025 at 2:32 PM
teeeeeeeee.bsky.social
`crossvalidation` and random search for calibrating support vector machines

https://thierrymoudiki.github.io/blog/2021/08/06/r/crossvalidation-svm-r

#Techtonique #DataScience #Python #rstats #MachineLearning
March 25, 2025 at 2:33 PM
teeeeeeeee.bsky.social
`crossvalidation` on R-universe, plus a classification example

https://thierrymoudiki.github.io/blog/2021/07/23/r/crossvalidation-r-universe

#Techtonique #DataScience #Python #rstats #MachineLearning
February 21, 2025 at 6:34 PM
timtriche.bsky.social
Maybe. Time to resurrect generalized crossvalidation?
October 9, 2023 at 1:38 AM
teeeeeeeee.bsky.social
Time series cross-validation using `crossvalidation` (Part 2)

https://thierrymoudiki.github.io/blog/2021/11/07/r/misc/timeseries/crossvalidation-eval-ts

#Techtonique #DataScience #Python #rstats #MachineLearning
Blog post image
April 24, 2025 at 10:11 PM
teeeeeeeee.bsky.social
parallel grid search cross-validation using `crossvalidation`

https://thierrymoudiki.github.io/blog/2021/07/31/r/parallel-crossvalidation

#Techtonique #DataScience #Python #rstats #MachineLearning
Blog post image
February 27, 2025 at 11:36 AM
teeeeeeeee.bsky.social
Time series cross-validation using `crossvalidation` (Part 2)

https://thierrymoudiki.github.io/blog/2021/11/07/r/misc/timeseries/crossvalidation-eval-ts

#Techtonique #DataScience #Python #rstats #MachineLearning
Blog post image
February 23, 2025 at 5:06 PM
weekend-editor.bsky.social
I spent a lot of my career learning that, for the applications in front of me, I didn't *need* much of the fancy statistics I loved so much.

Simple logistic regression and crossvalidation was enough, frighteningly often.

The classic methods are classic for a *reason*.
August 27, 2024 at 2:31 PM
teeeeeeeee.bsky.social
Comparing cross-validation results using crossval_ml and boxplots
https://thierrymoudiki.github.io/blog/2023/08/27/r/misc/crossvalidation-boxplots

#Python #DataScience #MachineLearning #rstats #Techtonique
Blog post image
August 2, 2026 at 1:25 PM
teeeeeeeee.bsky.social
Comparing cross-validation results using crossval_ml and boxplots

https://thierrymoudiki.github.io/blog/2023/08/27/r/misc/crossvalidation-boxplots

#Techtonique #DataScience #Python #rstats #MachineLearning
Blog post image
October 13, 2025 at 11:21 AM
kubernetes.activitypub.awakari.com.ap.brid.gy
Machine Learning Fundamentals: cross validation Cross Validation in Production Machine Learning Systems 1. Introduction In Q3 2023, a critical anomaly detection model powering fraud prevention at a...

#machinelearning #ai #crossvalidation

Origin | Interest | Match
Machine Learning Fundamentals: cross validation
Cross Validation in Production Machine Learning Systems 1. Introduction In Q3...
dev.to
July 9, 2025 at 5:06 PM
weekend-editor.bsky.social
To me, "Doing it right" meant: First get biologists to do adequately powered expts with enough controls. Then feature selection. Then logistic regression with LASSO regularization. All that under 5-fold crossvalidation. Then reduce to a gene set & re-test w/single sample gene set enrichment. 2/3
October 16, 2025 at 2:47 PM
byrdnick.com
Are #relationship rules (such as a 5:1 ratio of positive:negative #feedback) really predictive?

Not always.

Gottman et al's alleged predictors of #divorce often fail to predict:
😬 just 21% true positives? doi.org/10.1111/j.17...
☹️ failed replication? doi.org/10.1111/j.17...
Compared to "Gottman et al. (1998) and the other divorce prediction studies summarized earlier", "The analyses presented here included many predictor variables that can be measured with little error (e.g., education) and both our development and crossvalidation samples were randomly drawn from a single, representative data set. Both of these characteristics should enhance the robustness of the prediction equation. Yet, the accuracy of the equation still dropped markedly when applied to an independent subsample (and even further when the actual population baserate of early divorce was taken into account). An equation with an initial overall accuracy of 90% ended up with a positive predictive value of 21%." "We predicted that ...a higher ratio of positive to negative affect [at T1] would predict the couple remaining intact ...at T2 (Hypothesis b). Findings from the logistic regression analyses (Table 4), however, indicated that neither the total amount of positive affect nor the ratio of positive to negative affect for either men or women were significantly related to the relationship remaining intact."
May 21, 2026 at 11:03 AM
malcmanishere.bsky.social
I would like to thank @pipeshotgun on Twitter/X for their neat deep dive into something I hadn't thought about in a show I really, really love. They have some people to thank as well.

If this does well here, I'd be happy to repost the other threads they did on Steven Universe.
From May 31, 2025: I want to thank my mutuals, @/AstroNat2002 for suggesting this perception of Pearl and Steven's relationship to me in the first place, and @/bisroses for helping me gather quotes from the official art books and podcast. And to rosaq on tumblr as I used some of their (Cont'd) From May 31, 2025: (Cont'd) clip recompilations. I must admit, this thread it's not as flashed out as I'd like it to be, like I said before this is a theme that's outside my field and I wouldn't like to go too far on the subject risking the possibility of saying something really harmful. (Cont'd) @pipeshotgun · May 31 (Cont'd) So take everything with a pinch of salt, remember these threads are small exercises I like to give myself and this is just my opinion. From May 31, 2025: Sources: Cimsir, E & Akdogan, R. (2021). Childhood Emotional Incest Scale (CEIS): Development, Validation, CrossValidation, and Reliability. JOURNAL OF COUNSELING PSYCHOLOGY Villines, Z & Akers, R. (2022). What is emotional incest? Causes, effects and recovery. Medical News Today. Laderer, Ashley. (2024). Emotional Incest Doesn't Mean What You Think It Means. Charlie Health.
July 18, 2025 at 5:11 PM
teeeeeeeee.bsky.social
parallel grid search cross-validation using `crossvalidation`
https://thierrymoudiki.github.io/blog/2021/07/31/r/parallel-crossvalidation

#Python #DataScience #MachineLearning #rstats #Techtonique
Blog post image
September 14, 2026 at 8:03 PM
teeeeeeeee.bsky.social
`crossvalidation` and random search for calibrating support vector machines
https://thierrymoudiki.github.io/blog/2021/08/06/r/crossvalidation-svm-r

#Python #DataScience #MachineLearning #rstats #Techtonique
September 4, 2026 at 11:53 PM