@LifeOfMyMoon @rlmcelreath That's unreasonable; you never have perfect models, you make tradeoffs between accuracy about different things.
-
-
Replying to @davidmanheim
@LifeOfMyMoon@rlmcelreath If you need f(inputs, x, y, z) to optimize x, or y, your best model is different. (Like the optimizer's curse).2 replies 1 retweet 0 likes -
Replying to @davidmanheim
@davidmanheim@rlmcelreath Yes, now the system is larger. Your model is higher fidelity in one part than another. Still a representation.1 reply 0 retweets 1 like -
Replying to @LifeOfMyMoon
@davidmanheim@rlmcelreath A Taylor series is a model of a function; you've *chosen* a point to approximate locally, but still has veracity.1 reply 0 retweets 1 like -
Replying to @LifeOfMyMoon
@LifeOfMyMoon@rlmcelreath Right; those are explicitly different models; which model you use depends on purpose - neither is "false".1 reply 0 retweets 0 likes -
Replying to @davidmanheim
@LifeOfMyMoon@rlmcelreath Actually, the no free lunch theorem effectively says explicitly that no model is best for all questions.1 reply 0 retweets 1 like -
Replying to @davidmanheim
@davidmanheim@rlmcelreath Definitely agree here. There is a degeneracy of states: multiple models have the same veracity, different focus1 reply 0 retweets 1 like -
Replying to @LifeOfMyMoon
@davidmanheim@rlmcelreath Except, NFL doesn't really apply: they hold information constant. Can form better model by obtaining perfect info1 reply 0 retweets 0 likes -
Replying to @LifeOfMyMoon
@LifeOfMyMoon@rlmcelreath Yes. Unfortunately, models exist to represent imperfectly, usually given available data.2 replies 0 retweets 0 likes -
@davidmanheim @rlmcelreath Models can be perfect or imperfect; unfortunately, hypothesis space for imperfect is much larger
-
-
Replying to @LifeOfMyMoon
@LifeOfMyMoon@rlmcelreath The space of perfect models isn't just of measure 0 - it's finite, and exactly equal to 1.1 reply 0 retweets 0 likes -
Replying to @davidmanheim
@davidmanheim@rlmcelreath Are you outlawing submodels and models of subsystems?1 reply 0 retweets 0 likes - 2 more replies
New conversation -
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.