You can use anything you’d like, but be aware that you should not take the mean of the posterior and compute a single statistic. You should compute the statistic for each posterior sample, then report the distribution over r2/mape/whatever
Related topics
| Topic | Replies | Views | Activity | |
|---|---|---|---|---|
| WAIC and R^2 scores comparison | 0 | 465 | May 9, 2019 | |
| Bayesian linear regression | 3 | 729 | October 31, 2019 | |
| Appropriate params for NegativeBinomial Likelihood and model evaluation | 1 | 589 | October 5, 2021 | |
| Comparing posterior predictive errors across samples | 0 | 404 | October 7, 2020 | |
| Predictions and evaluation of linear regression | 11 | 1866 | November 22, 2018 |