@cdunnpasadena
That is, in a Bayesian context there's some probabilities for unknown inputs, there's some predictor apparatus, and there's some set of probabilities for outcomes. If the observed outcomes don't agree with the predicted ones it's either because your assumptions are wrong about the unknown inputs, or your predictor is wrong, it's not because "probability doesn't work"
=> More informations about this toot | View the thread | More toots from dlakelan@mastodon.sdf.org
=> View cdunnpasadena@sfba.social profile
text/geminiThis content has been proxied by September (3851b).