Aki Vehtari
Aki Vehtari
New PSIS version based on updated https://arxiv.org/abs/1507.02646v5
From Lu Cheng: "It was said in GPStuff manual page 42 that periodic kernel was coming from this paper http://jmlr.org/proceedings/papers/v33/solin14.pdf In page 907, equation (23) and GPStuff appendix, there is...
Continuous Ranked Probability Score could be useful alternative for log score, As it is comparing predictive distribution to the observed data, it's a bit more complicated to implement than point...
Example output from https://discourse.mc-stan.org/t/understanding-loo-results-for-non-standard-contaminated-controls-model/25493 ``` Pareto k diagnostic values: Count Pct. Min. n_eff (-Inf, 0.5] (good) 717 73.2% 151 (0.5, 0.7] (ok) 133 13.6% 116 (0.7, 1] (bad) 74 7.6%...
The current output ``` Pareto k diagnostic values: Count Pct. Min. n_eff (-Inf, 0.5] (good) 717 73.2% 151 (0.5, 0.7] (ok) 133 13.6% 116 (0.7, 1] (bad) 74 7.6% 38...
A few times users have asked how much variability there can be in khat estimate if they would re-run MCMC. We can estimate that by looking at the marginal posterior...
The loo repo has only GPL-3 license. @aloctavodia asked to re-use LOO glossary. Maybe we should add also explicit CC license for the text content? Stan manual is with CC-BY...
If the list of loo objects has some named elements the output column names behave differently. For example, ``` > loo_compare(list(loo1,`Model f1+f2`=loo2)) elpd_diff se_diff Model f1+f2 0.0 0.0 model1 -530.0...
Section 4.3 of https://avehtari.github.io/modelselection/diabetes.html demonstrates how to use loo package with other loss/utility function than log score. LOO-R^2 is another example. Both of these are just for point predictions, which...
The package is named as loo, but when psis-loo fails and k-fold-cv is used, then corresponding function could form a object which could be given to compare function.