{
  "$type": "site.standard.document",
  "bskyPostRef": {
    "cid": "bafyreidn3xvnzhydyyzxo6xr3ai5lzbh3eeth5ssdsgrs5evmqtb3xyc6m",
    "uri": "at://did:plc:wwyqal4cnqhuwyacdj7rqq3n/app.bsky.feed.post/3mhuktn6ofq72"
  },
  "path": "/t/internal-validation-with-bayesian-models/28673#post_1",
  "publishedAt": "2026-03-25T00:22:27.000Z",
  "site": "https://discourse.datamethods.org",
  "textContent": "In the clinical prediction model framework, many suggest doing internal validation after model development to quantify and adjust for optimism/overfitting. The most recommended procedure is doing non-parametric bootstrapping 200-500 times.\n\nHowever, doing such a procedure with a Bayesian model and a large dataset is infeasible due to computational burden. In the Bayesian literature, model checking with posterior predictive checks and even leave-one-out cross-validation are commonly recommended. Yet I have not yet seen these procedures in the clinical prediction literature.\n\nHow should one internally validate a Bayesian clinical prediction model?",
  "title": "Internal Validation with Bayesian models"
}