# \#projpred

**URL:** https://discourse.mc-stan.org/tag/projpred/136.md

[Latest](https://discourse.mc-stan.org/latest.md) · [Categories](https://discourse.mc-stan.org/categories.md) · [Tags](https://discourse.mc-stan.org/tags.md)

---

## [Projpred interpretation](https://discourse.mc-stan.org/t/projpred-interpretation/41223)

<div class="topic-metadata">

**Author:** [@ViktorVdV](https://discourse.mc-stan.org/u/ViktorVdV)\
**Replies:** 2\
**Last updated:** [May 16, 2026, 11:42am UTC](https://discourse.mc-stan.org/t/projpred-interpretation/41223 "2026-05-16T11:42:50Z")

</div>

Hi all, I am using projpred to reduce my predictor set. One of my colleagues raised a concern that I would like to get some expert opinion on. The concern is about correlated predictors and the risk of over-interpretin…

---

## [Using projpred with a hierarchical Stan model with posterior-draw measurement error](https://discourse.mc-stan.org/t/using-projpred-with-a-hierarchical-stan-model-with-posterior-draw-measurement-error/41082)

<div class="topic-metadata">

**Author:** [@ViktorVdV](https://discourse.mc-stan.org/u/ViktorVdV)\
**Replies:** 3\
**Last updated:** [May 7, 2026, 9:30am UTC](https://discourse.mc-stan.org/t/using-projpred-with-a-hierarchical-stan-model-with-posterior-draw-measurement-error/41082 "2026-05-07T09:30:20Z")

</div>

Hi all, I’m working with a Stan model where I explicitly propagate uncertainty from a previous model by using posterior draws as inputs. I’m now trying to perform predictor selection among 9 covariates that exhibit some…

---

## [Implementing projection predictive feature selection for custom model with \`projpred\`](https://discourse.mc-stan.org/t/implementing-projection-predictive-feature-selection-for-custom-model-with-projpred/40885)

<div class="topic-metadata">

**Author:** [@helske](https://discourse.mc-stan.org/u/helske)\
**Replies:** 6\
**Last updated:** [February 6, 2026, 8:27pm UTC](https://discourse.mc-stan.org/t/implementing-projection-predictive-feature-selection-for-custom-model-with-projpred/40885 "2026-02-06T20:27:36Z")

</div>

I’m trying to familiarize myself with the projection predictive approach and projpred package, and thinking about implementing projpred methods for particular custom model class. However, the documentation on how to do t…

---

## [Using {projpred} latent projection with {brms} Weibull family models](https://discourse.mc-stan.org/t/using-projpred-latent-projection-with-brms-weibull-family-models/39275)

<div class="topic-metadata">

**Author:** [@rtnliqry](https://discourse.mc-stan.org/u/rtnliqry)\
**Replies:** 31\
**Last updated:** [November 26, 2025, 10:29pm UTC](https://discourse.mc-stan.org/t/using-projpred-latent-projection-with-brms-weibull-family-models/39275 "2025-11-26T22:29:56Z")

</div>

Hi Stan Forum! I am currently trying to implement the use of the excellent projpred package (thanks @fweber144 and colleagues!) for variable selection with brms parametric survival models that use a Weibull response dis…

---

## [Projpred: Fixing Group Effects in Search Terms and Tips for Speed?](https://discourse.mc-stan.org/t/projpred-fixing-group-effects-in-search-terms-and-tips-for-speed/31678)

<div class="topic-metadata">

**Author:** [@l-gorman](https://discourse.mc-stan.org/u/l-gorman)\
**Replies:** 7\
**Last updated:** [September 3, 2025, 10:16am UTC](https://discourse.mc-stan.org/t/projpred-fixing-group-effects-in-search-terms-and-tips-for-speed/31678 "2025-09-03T10:16:08Z")

</div>

Firstly, thanks so much for all of the work on BRMS, STAN, and projpred, these are all fantastic tools!!! And apologies in advance if my question is daft, I am still learning about these tools and I have likely overlooke…

---

## [Usage and Interpreation of cv\_varsel-function](https://discourse.mc-stan.org/t/usage-and-interpreation-of-cv-varsel-function/38528)

<div class="topic-metadata">

**Author:** [@ria](https://discourse.mc-stan.org/u/ria)\
**Replies:** 13\
**Last updated:** [August 21, 2025, 1:41pm UTC](https://discourse.mc-stan.org/t/usage-and-interpreation-of-cv-varsel-function/38528 "2025-08-21T13:41:25Z")

</div>

@fweber144 I wanted to thank you for the informative video you shared about how to use the projpred package. I found it very helpful. Moreover I did read the following homage, which includes an example and R-Code: Projec…

---

## [New projpred version (2.9.0) on CRAN](https://discourse.mc-stan.org/t/new-projpred-version-2-9-0-on-cran/39862)

<div class="topic-metadata">

**Author:** [@fweber144](https://discourse.mc-stan.org/u/fweber144)\
**Replies:** 2\
**Last updated:** [July 21, 2025, 6:51pm UTC](https://discourse.mc-stan.org/t/new-projpred-version-2-9-0-on-cran/39862 "2025-07-21T18:51:28Z")

</div>

A new projpred version (2.9.0) has been released on CRAN. The most important changes are: Subsampled PSIS-LOO CV (see argument nloo of cv\_varsel()) has been fixed and is not experimental anymore. Thanks to @avehtari fo…

---

## [Projpred: latent projection for hierarchical ordinal models](https://discourse.mc-stan.org/t/projpred-latent-projection-for-hierarchical-ordinal-models/39779)

<div class="topic-metadata">

**Author:** [@melex](https://discourse.mc-stan.org/u/melex)\
**Replies:** 8\
**Last updated:** [July 9, 2025, 7:28am UTC](https://discourse.mc-stan.org/t/projpred-latent-projection-for-hierarchical-ordinal-models/39779 "2025-07-09T07:28:39Z")

</div>

Hi everyone, I am back with another question on projected variable selection for hierarchical models. The developers kindly covered in detail different aspects of categorical models here. In the linked post, @fweber144 …

---

## [Projpred: selection of submodel that doesn't replicate predictive performance of reference model](https://discourse.mc-stan.org/t/projpred-selection-of-submodel-that-doesnt-replicate-predictive-performance-of-reference-model/39482)

<div class="topic-metadata">

**Author:** [@melex](https://discourse.mc-stan.org/u/melex)\
**Replies:** 13\
**Last updated:** [June 2, 2025, 11:06pm UTC](https://discourse.mc-stan.org/t/projpred-selection-of-submodel-that-doesnt-replicate-predictive-performance-of-reference-model/39482 "2025-06-02T23:06:23Z")

</div>

Hi everyone, First of all, thank you to the developers of this great package - Code / Program outputprojpredCode / Program output is great, as is all of the support and materials provided. I would like to ask your opin…

---

## [Projpred: modify labels in plot.vsel](https://discourse.mc-stan.org/t/projpred-modify-labels-in-plot-vsel/39594)

<div class="topic-metadata">

**Author:** [@melex](https://discourse.mc-stan.org/u/melex)\
**Replies:** 5\
**Last updated:** [May 30, 2025, 3:23am UTC](https://discourse.mc-stan.org/t/projpred-modify-labels-in-plot-vsel/39594 "2025-05-30T03:23:38Z")

</div>

Hi everyone, Back again with another query on projpred - this time about modifying the plots produced by plot(\<cv\_proportions\>), plot(\<ranking\>), and plot(\<vsel\>). I would like to have more control over the plots, for…

---

## [Projpred: augmented and latent projection for hierarchical categorical models](https://discourse.mc-stan.org/t/projpred-augmented-and-latent-projection-for-hierarchical-categorical-models/39537)

<div class="topic-metadata">

**Author:** [@melex](https://discourse.mc-stan.org/u/melex)\
**Replies:** 3\
**Last updated:** [May 26, 2025, 5:36am UTC](https://discourse.mc-stan.org/t/projpred-augmented-and-latent-projection-for-hierarchical-categorical-models/39537 "2025-05-26T05:36:03Z")

</div>

Hi everyone, First of all, thanks again to everyone, especially the developers, in this community for all their support. Projpred is such a great package and I’m excited to explore it more. I have a hierarchical catego…

---

## [Mac throttling CPU with kernel\_task](https://discourse.mc-stan.org/t/mac-throttling-cpu-with-kernel-task/39497)

<div class="topic-metadata">

**Author:** [@lomez](https://discourse.mc-stan.org/u/lomez)\
**Replies:** 5\
**Last updated:** [May 16, 2025, 9:48pm UTC](https://discourse.mc-stan.org/t/mac-throttling-cpu-with-kernel-task/39497 "2025-05-16T21:48:41Z")

</div>

Has anyone encountered CPU throttling issues on a Mac? I am running projpred’s cv\_varsel with LOO and after running for maybe 18-24 hours, kernel\_task basically shuts down all of my cores to manage the temperature, and e…

---

## [Could exploitation based adaptive sampling be performed before using Projection Predictive variable selection?](https://discourse.mc-stan.org/t/could-exploitation-based-adaptive-sampling-be-performed-before-using-projection-predictive-variable-selection/38482)

<div class="topic-metadata">

**Author:** [@saya](https://discourse.mc-stan.org/u/saya)\
**Replies:** 4\
**Last updated:** [January 15, 2025, 5:42pm UTC](https://discourse.mc-stan.org/t/could-exploitation-based-adaptive-sampling-be-performed-before-using-projection-predictive-variable-selection/38482 "2025-01-15T17:42:19Z")

</div>

Hello, I have been having this question in mind for so long now, and couldn’t manage to find a clear answer. If I am going to use projection predictive as a sensitivity analysis tool since I am interested in finding out…

---

## [Projpred error](https://discourse.mc-stan.org/t/projpred-error/37314)

<div class="topic-metadata">

**Author:** [@ssalimi](https://discourse.mc-stan.org/u/ssalimi)\
**Replies:** 7\
**Last updated:** [November 28, 2024, 1:40am UTC](https://discourse.mc-stan.org/t/projpred-error/37314 "2024-11-28T01:40:14Z")

</div>

I ran a model for 144 features for 749 sample size (N=749). I fit the model using rstanarm. D \<- ncol(df3\[, 15:158\]) # Prior guess p0 \<- 5 # Number of observations: N \<- nrow(df3) tau0 \<- p0 / (D - p0) \* 1 / s…

---

## [Projpred error: "factor has new levels"](https://discourse.mc-stan.org/t/projpred-error-factor-has-new-levels/36873)

<div class="topic-metadata">

**Author:** [@vvp1](https://discourse.mc-stan.org/u/vvp1)\
**Replies:** 6\
**Last updated:** [November 7, 2024, 3:14pm UTC](https://discourse.mc-stan.org/t/projpred-error-factor-has-new-levels/36873 "2024-11-07T15:14:44Z")

</div>

I have fitted a logistic regression with a regularised horshoe prior using rstanarm. The number of covariates is over 7000 and a few of them are categorical. One of the categorical covariates is race and has tree levels…

---

## [Errors running cv\_varsel , and a search terms question](https://discourse.mc-stan.org/t/errors-running-cv-varsel-and-a-search-terms-question/31686)

<div class="topic-metadata">

**Author:** [@M-J-Mitchell](https://discourse.mc-stan.org/u/M-J-Mitchell)\
**Replies:** 14\
**Last updated:** [June 23, 2024, 3:39pm UTC](https://discourse.mc-stan.org/t/errors-running-cv-varsel-and-a-search-terms-question/31686 "2024-06-23T15:39:54Z")

</div>

Hi all, I am a PhD student using brms to run mixed models to understand the effect of parental factors on the fertility of eggs. My model currently looks like this - understanding the effect ion maternal and paternal ag…

---

## [Help interpreting projpred's predictive performance plot](https://discourse.mc-stan.org/t/help-interpreting-projpreds-predictive-performance-plot/35026)

<div class="topic-metadata">

**Author:** [@sayianka](https://discourse.mc-stan.org/u/sayianka)\
**Replies:** 5\
**Last updated:** [May 22, 2024, 9:30pm UTC](https://discourse.mc-stan.org/t/help-interpreting-projpreds-predictive-performance-plot/35026 "2024-05-22T21:30:57Z")

</div>

I am fitting a model to proportion data (working on neonatal mortality), and the proportions contain a lot of zeros, so I chose the zero-inflated beta. Here is a density plot of my outcome variable: The model is: m…

---

## [Projpred results interpretation](https://discourse.mc-stan.org/t/projpred-results-interpretation/34974)

<div class="topic-metadata">

**Author:** [@Iain\_Gallagher](https://discourse.mc-stan.org/u/Iain_Gallagher)\
**Replies:** 2\
**Last updated:** [May 1, 2024, 4:33pm UTC](https://discourse.mc-stan.org/t/projpred-results-interpretation/34974 "2024-05-01T16:33:28Z")

</div>

We’re working on a project that examines whether a set of variables used in the clinic plus some metabolites is capable of tracking changes in another clinical variable over an intervention. We have created our model and…

---

## [How are sub-models fitted in Projection predictive inference (projpred)?](https://discourse.mc-stan.org/t/how-are-sub-models-fitted-in-projection-predictive-inference-projpred/34511)

<div class="topic-metadata">

**Author:** [@sayianka](https://discourse.mc-stan.org/u/sayianka)\
**Replies:** 4\
**Last updated:** [March 20, 2024, 7:37am UTC](https://discourse.mc-stan.org/t/how-are-sub-models-fitted-in-projection-predictive-inference-projpred/34511 "2024-03-20T07:37:46Z")

</div>

@fweber144 I have been using the excellent projpred package to conduct variable selection on fitted models from brms. I have also read a few papers on projection predictive inference . Assuming I fit a simple linear r…

---

## [Variable selection with cmdstanr and projpred](https://discourse.mc-stan.org/t/variable-selection-with-cmdstanr-and-projpred/34435)

<div class="topic-metadata">

**Author:** [@fabio](https://discourse.mc-stan.org/u/fabio)\
**Replies:** 4\
**Last updated:** [March 11, 2024, 6:44pm UTC](https://discourse.mc-stan.org/t/variable-selection-with-cmdstanr-and-projpred/34435 "2024-03-11T18:44:58Z")

</div>

Hi dear Stan community! I would like to perform Projection Predictive Feature Selection - more specifically the rankig of the predictors - and using the cmdstanr interface to handle R code, Stan code, compliations… Wh…

---

## [New projpred version (2.8.0) on CRAN](https://discourse.mc-stan.org/t/new-projpred-version-2-8-0-on-cran/33593)

<div class="topic-metadata">

**Author:** [@fweber144](https://discourse.mc-stan.org/u/fweber144)\
**Replies:** 0\
**Last updated:** [December 15, 2023, 9:33am UTC](https://discourse.mc-stan.org/t/new-projpred-version-2-8-0-on-cran/33593 "2023-12-15T09:33:42Z")

</div>

A new projpred version (2.8.0) has been released on CRAN. Most important changes are: Search results generated in an earlier varsel() or cv\_varsel() call can now be re-used by the help of the new varsel.vsel() and cv\_v…

---

## [New projpred version (2.7.0) on CRAN](https://discourse.mc-stan.org/t/new-projpred-version-2-7-0-on-cran/32946)

<div class="topic-metadata">

**Author:** [@fweber144](https://discourse.mc-stan.org/u/fweber144)\
**Replies:** 1\
**Last updated:** [October 2, 2023, 7:24am UTC](https://discourse.mc-stan.org/t/new-projpred-version-2-7-0-on-cran/32946 "2023-10-02T07:24:25Z")

</div>

A new projpred version (2.7.0) has been released on CRAN. Most noteworthy changes are that the default search method is forward search now, that a built-in connection to the posterior package has been added for handling …

---

## [CV Varsel Error: Infinite or missing values in 'x'](https://discourse.mc-stan.org/t/cv-varsel-error-infinite-or-missing-values-in-x/31703)

<div class="topic-metadata">

**Author:** [@l-gorman](https://discourse.mc-stan.org/u/l-gorman)\
**Replies:** 12\
**Last updated:** [September 6, 2023, 7:46pm UTC](https://discourse.mc-stan.org/t/cv-varsel-error-infinite-or-missing-values-in-x/31703 "2023-09-06T19:46:22Z")

</div>

Hi again! So… apologies for yet another query, and thanks for all the help you’ve provided before. I am running a hierarchical model with 30,000 observations. I am using projpred for variable selection, more details on …

---

## [Projpred: Non-Convergence of Predictive Performance with Reference Model](https://discourse.mc-stan.org/t/projpred-non-convergence-of-predictive-performance-with-reference-model/31916)

<div class="topic-metadata">

**Author:** [@l-gorman](https://discourse.mc-stan.org/u/l-gorman)\
**Replies:** 20\
**Last updated:** [August 18, 2023, 11:34am UTC](https://discourse.mc-stan.org/t/projpred-non-convergence-of-predictive-performance-with-reference-model/31916 "2023-08-18T11:34:07Z")

</div>

Hi again! Apologies for yet another query, I seem to be a boomerang of projpred queries. Thanks so much for all the help so far @fweber144. It’s incredible to see such an active community! I have been trying to fit a m…

---

## [Interpretation of CIs of acc statistic](https://discourse.mc-stan.org/t/interpretation-of-cis-of-acc-statistic/32441)

<div class="topic-metadata">

**Author:** [@T\_nick](https://discourse.mc-stan.org/u/T_nick)\
**Replies:** 4\
**Last updated:** [August 15, 2023, 9:52am UTC](https://discourse.mc-stan.org/t/interpretation-of-cis-of-acc-statistic/32441 "2023-08-15T09:52:45Z")

</div>

Can anyone provide any information about how the acc statistic and the corresponding CIs are calculated during the variable selection procedure with projpred? Many thanks

---

## [Projpred workflow paper out now](https://discourse.mc-stan.org/t/projpred-workflow-paper-out-now/31895)

<div class="topic-metadata">

**Author:** [@fweber144](https://discourse.mc-stan.org/u/fweber144)\
**Replies:** 0\
**Last updated:** [June 28, 2023, 11:17am UTC](https://discourse.mc-stan.org/t/projpred-workflow-paper-out-now/31895 "2023-06-28T11:17:55Z")

</div>

A new projpred paper by Yann McLatchie, Sölvi Rögnvaldsson, me, and @avehtari has been published on arXiv: \[2306.15581\] Robust and efficient projection predictive inference. The main purpose of it is to present a workfl…

---

## [Projpred with multiple imputations (brm\_multiple) - does this example work?](https://discourse.mc-stan.org/t/projpred-with-multiple-imputations-brm-multiple-does-this-example-work/31456)

<div class="topic-metadata">

**Author:** [@Andrew\_Y](https://discourse.mc-stan.org/u/Andrew_Y)\
**Replies:** 3\
**Last updated:** [June 15, 2023, 12:00pm UTC](https://discourse.mc-stan.org/t/projpred-with-multiple-imputations-brm-multiple-does-this-example-work/31456 "2023-06-15T12:00:17Z")

</div>

It seems that using projpred (and loo) with multiple imputation isn’t fully worked out (?), but there have been some useful suggestions (mainly by @avehtari) on these forums, e.g., Projpred - model with horseshoe prior d…

---

## [New projpred version (2.6.0) on CRAN](https://discourse.mc-stan.org/t/new-projpred-version-2-6-0-on-cran/31585)

<div class="topic-metadata">

**Author:** [@fweber144](https://discourse.mc-stan.org/u/fweber144)\
**Replies:** 0\
**Last updated:** [June 1, 2023, 11:22am UTC](https://discourse.mc-stan.org/t/new-projpred-version-2-6-0-on-cran/31585 "2023-06-01T11:22:52Z")

</div>

A new projpred version (2.6.0) has been released on CRAN. Most noteworthy user-facing changes are the new functions ranking(), cv\_proportions(), and plot.cv\_proportions() as well as a major update of plot.vsel(). See the…

---

## [Projpred 2.5.0](https://discourse.mc-stan.org/t/projpred-2-5-0/30989)

<div class="topic-metadata">

**Author:** [@fweber144](https://discourse.mc-stan.org/u/fweber144)\
**Replies:** 0\
**Last updated:** [April 6, 2023, 7:01am UTC](https://discourse.mc-stan.org/t/projpred-2-5-0/30989 "2023-04-06T07:01:15Z")

</div>

A new projpred version (2.5.0) has been released on CRAN. Most noteworthy user-facing changes are an enhanced verbose mode (extensible by a new global option, thanks to @jtimonen), a revised “Troubleshooting” section in …

---

## [I cannot seem to make varsel() work from brms](https://discourse.mc-stan.org/t/i-cannot-seem-to-make-varsel-work-from-brms/30612)

<div class="topic-metadata">

**Author:** [@egumasa](https://discourse.mc-stan.org/u/egumasa)\
**Replies:** 2\
**Last updated:** [March 9, 2023, 4:02am UTC](https://discourse.mc-stan.org/t/i-cannot-seem-to-make-varsel-work-from-brms/30612 "2023-03-09T04:02:01Z")

</div>

Hi, I am trying to reproduce the example on how to use varsel and cv\_varsel(), and it does not seem to work on my end. I was not sure if this was a bug or I am missing something, so I am posting here before creating an …

[Next page](https://discourse.mc-stan.org/tag/projpred/136.md?match_all_tags=true&page=1&tags%5B%5D=projpred)