{
  "annotation_id": "9b996c88-c35f-49ff-8e31-690bb43f90b2",
  "date_created": "2026-03-02T18:00:47.043000Z",
  "date_modified": "2026-03-02T18:00:47.043000Z",
  "file_hash": "48b944f2d82375d739f48cd17dbb179d365ec9110b24cb030b81afc43d7cad08",
  "private": false,
  "record": {
    "abstract": "The problem of assigning probability distributions which objectively reflect\nthe prior information available about experiments is one of the major stumbling\nblocks in the use of Bayesian methods of data analysis. In this paper the\nmethod of Maximum (relative) Entropy (ME) is used to translate the information\ncontained in the known form of the likelihood into a prior distribution for\nBayesian inference. The argument is inspired and guided by intuition gained\nfrom the successful use of ME methods in statistical mechanics. For experiments\nthat cannot be repeated the resulting \"entropic prior\" is formally identical\nwith the Einstein fluctuation formula. For repeatable experiments, however, the\nexpected value of the entropy of the likelihood turns out to be relevant\ninformation that must be included in the analysis. The important case of a\nGaussian likelihood is treated in detail.",
    "arxiv_id": "physics/0307055",
    "authors": [
      "Ariel Caticha",
      "Roland Preuss"
    ],
    "categories": [
      "physics.data-an"
    ],
    "doi": "10.1103/PhysRevE.70.046127",
    "journal_ref": "Phys.Rev. E70 (2004) 046127",
    "title": "Maximum Entropy and Bayesian Data Analysis: Entropic Priors",
    "url": "https://arxiv.org/abs/physics/0307055"
  },
  "schema_id": "dorsal/arxiv",
  "source": {
    "execution_id": "fef18981-5fd5-4ced-b7b8-9f536614d1be",
    "id": "arXiv Dataset IDs",
    "type": "Model",
    "variant": "snapshot-2026-03-01",
    "version": "0.1.0"
  },
  "user_id": 1000002
}