Timeline for Elements of Statistical Learning - Statistical Decision Theory : Doubt regarding Minimization of EPE
Current License: CC BY-SA 3.0
6 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Aug 14, 2022 at 14:01 | comment | added | newandlost | I still don't understand how to get from 2.11 to 2.12, $E_X$ is not a constant, if it the expectation value of everything the follows to the right of that symbol if I understand it correctly. I mean expression 2.11 is $\int dx p(x) \int dy (y-f(x))^2 p(y|x)$ , so how come not to considere the integration over $x$? | |
| Aug 14, 2022 at 13:53 | comment | added | newandlost | I still don't fully understand it: Assuming $E_{Y|X} ([Y-c]^2|X=x) = \int dy (y^2+c^2-2yc)p(y|x)$, we can get the minimum by taking the derivative, setting to zero and solving c yields: $c = \frac{E_{Y|X} (Y|X=x) }{\int dy\; p(y|x)}$ -> ??? Where is my mistake? | |
| Jul 2, 2017 at 11:31 | comment | added | David Epstein | @usεr11852 Not that good any more (too old) | |
| Jul 2, 2017 at 11:29 | history | edited | David Epstein | CC BY-SA 3.0 |
improved notation at one point
|
| Jul 2, 2017 at 0:26 | comment | added | usεr11852 | (+1) Totally unrelated but I have to ask based on your location: Are you "really, really good" in Geometry? | |
| Jul 1, 2017 at 20:31 | history | answered | David Epstein | CC BY-SA 3.0 |