SBC
SBC copied to clipboard
hyunjimoon
The docs at https://mc-stan.org/rstan/reference/sbc.html discuss how such model looks like. Unfortunately, I don't think there is an easy way to easily plug them into the existing architecture, as models for...
- Why priors are important - Why one may want different priors for SBC than for real model - Discuss rejection sampling (details already presented in the `rejection_sampling` vignette)
S = 10 ends within 1 minute, but with 100 it doesn't seem to finish (for over an hour) in `result_tn_100
Code is much too long to post here (but can send via email if needed) but I _sometimes_ get this error when running many replicates (1000): ``` Error in if...
using pvalue as a global summary could be misleading as the following example suggests: ``` a
Wholistic measure on quantile, a summary statistics for ecdf needeed. - [Bayesian Quantile Matching Estimation](https://arxiv.org/pdf/2008.06423.pdf)
1. Could we add some indication (prior or post) to`simulated_values` name? I was confused for a long time :D It seems likely that `simulated_value` are prior sample values from [this](https://github.com/hyunjimoon/SBC/blob/cc189fafd8bbb1a36797f4fb9b9af61f255ab458/R/results.R#L430)...
@rybern and I are working on a model network navigator (MNN) and this experiment explores the possibility of defining model correlation metrics and how to apply them in MNN.
Is there any way to not get data as an input for inference engine [here](https://github.com/hyunjimoon/SBC/blob/e31813bce3e9e5160bc2e1c40d59f982efbfb3e7/R/sample.R#L204)?
What? find the self-consistency parmater regions via contraction mapping Assumption: variance of parameter samples would be similar before and after in self-consistency parmater regions How? run until the ratio <...