-
I just discovered `crps`! This is great, I was actually asking about how to compare probabilistic vs non-probabilistic predictions https://discourse.mc-stan.org/t/comparison-of-machine-learning-models…
-
Currently, mcse_elpd_loo is computed by assuming the pointwise mcse's to be independent. As the same MC draws are used for each pointwise PSIS weights, the weights are correlated and this correlation …
-
### Describe the issue
[Issue]: Write samle base on SemanticKernel with content copy from Example04_Dynamic_GroupChat_Coding_Task, can't work. I don't know what's wrong happend.
This is my code : h…
-
A plot where each row corresponds to study/group, each row showing (1) interval of hypermean in a LOO model, (2) point estimate (and maybe interval) for that group
-
2023-05-15 03:05:14 [INFO] ---------------------------------------------
2023-05-15 03:05:14 [INFO] --- GitHub Actions Multi Language Linter ----
2023-05-15 03:05:14 [INFO] - Image Creation Dat…
-
There are several useful functions that can be used for model assessment and comparison that could be added here. In some cases these are already available in [loo](https://github.com/stan-dev/loo) or…
-
Coffee Spots in Seattle
Seattle Essential Coffee Spots!
This is my first headline 2!
There are many coffee places in Seattle. Each co…
-
Now that we've changed over to posterior rather than loo for the psis code, we could think about submitting to CRAN so that other packages could use the iwmm functionality. Specifically I'm thinking o…
-
### 🎮 Game Request
it will ask for ingredients any wrong pick all points loos and it will say wrong serve!
### Point down the features
Game points:
Correct combination of food items will add 10 ma…
-
See https://dictionary.cambridge.org/us/dictionary/english/loo
Is there some magnificent joke that I'm missing? Otherwise, hardly seem worth it. Do we already have other candidate names?