🗻Surfacing assumption iceberg

[hyunjimoon]

new paper on SBC which I wish to reboot with the name behavior test

a. Paper introduce "Theory and Practice of Choosing Test Quantities for SBC"

b. suggested behavior test as SBC's new name (mail below)

c. customized prior and resolution i.e. polynomial function approximation of prior distribution based on user input CDF (current grant i.e. probability distributions are not made for priors) and setting time step for desired resolution are two problem needs to be solved. From algorithm in #11 for pattern classification for reference modes, segment is the key but setting it is vague #19 i.e. importance of hyperparameter; the same problem in simulation-based test.
d. simulation-based test's goal is inferential behavioral of model over an ensemble of possible assumptions and observations

e. modelers are giving stimulus to test organism (model)'s response (behavior := actions and mannerisms made by individuals, organisms, systems in some environment. ... computed response of the system or organism to various stimuli or inputs)

f. how to herald the part underneath? Assumed part of the model (e.g. approximator's Hessian matrix, level of hierarchy (gardens of forking path; we see what we expect to see), hyperparameter or hyperprior) that should be tested

[hyunjimoon]

Stanify classified icebergs into four types: precision (S, M, N, P, PH, Q, R), setting, numeric, prior.

[hyunjimoon]

revisited this thread as i was investigating the connection between the subjective prior with jeff in relation with rational meaning construction (RMC). we defined RMC to be Action oriented encoding of belief. where the probabilistic language of thought