Dominique Makowski

I wonder if that formula: https://github.com/easystats/bayestestR/blob/6337d43e566ad211b97a5abf344aa07f99a8ca60/R/estimate_density.R#L326-L334 makes sense with other estimation l, like KernSmooth or logspline @mattansb

I suppose the assumption of normality is regarding each value, like the density at each value is normally distributed ^^ but yeah idk man I'm just doing copypasta 😁 but...

The problem is that it can lead to negative densities for the CI_low 😕 ``` r tail(bayestestR::estimate_density(iris$Sepal.Width)) #> x y CI_low CI_high #> 1019 4.388270 0.02951812 -0.004070569 0.06310681 #> 1020...

I'mma put the default to NULL so that it doens't return the CI by default, as it's rather experimental 😁

Sounds like a good idea

@strengejacke I think it'd be good to have an `effects` and `components` argument for describe_priors, following the standards of describe_posteriors (so that we can match them easily). However, I wonder...

Right I see, I was trying to fix this bug in the checks before CRAN submission: ```r ── Failure (test-describe_prior.R:85:5): describe_prior ──────────────────────── describe_prior(mod_brms) (`actual`) not equal to structure(...) (`expected`). `attr(actual,...

Note to myself: add a paragraph about aliases

I think the time has come, before we push to CRAN., I'll find some time to do that this WE / beginning next week (and finally check that SI function...

``` r model # R2 for Mixed Models #> #> Conditional R2: 0.703 #> Marginal R2: 0.703 performance::r2(model$gam) #> $R2 #> Adjusted R2 #> 0.7371402 ``` Created on 2021-02-16 by...