Pythagorean theorem in linear regression using matrix notation
$begingroup$
I am reading online scribe notes on linear regression and I am being quite confused (since did not do linear algebra for a while) on how Pythagorean theorem is being applied here:
First of all I am confused of what "cost" is and then how we get from
$$arg min_x ||A(x-x') - bb'||_2^2 $$ to
$$arg min_x ||A(x-x')||_2^2 - ||b'||_2^2 $$ ? E.g I need some intuition and mathematical explanation of what is happening here and how Pythagorean theorem is used in vector space.
Also why $$arg min_x ||Ax - b||_2 equiv arg min_x ||Ax - b||_2^2 $$?
Thank you!
linear-algebra regression
$endgroup$
add a comment |
$begingroup$
I am reading online scribe notes on linear regression and I am being quite confused (since did not do linear algebra for a while) on how Pythagorean theorem is being applied here:
First of all I am confused of what "cost" is and then how we get from
$$arg min_x ||A(x-x') - bb'||_2^2 $$ to
$$arg min_x ||A(x-x')||_2^2 - ||b'||_2^2 $$ ? E.g I need some intuition and mathematical explanation of what is happening here and how Pythagorean theorem is used in vector space.
Also why $$arg min_x ||Ax - b||_2 equiv arg min_x ||Ax - b||_2^2 $$?
Thank you!
linear-algebra regression
$endgroup$
add a comment |
$begingroup$
I am reading online scribe notes on linear regression and I am being quite confused (since did not do linear algebra for a while) on how Pythagorean theorem is being applied here:
First of all I am confused of what "cost" is and then how we get from
$$arg min_x ||A(x-x') - bb'||_2^2 $$ to
$$arg min_x ||A(x-x')||_2^2 - ||b'||_2^2 $$ ? E.g I need some intuition and mathematical explanation of what is happening here and how Pythagorean theorem is used in vector space.
Also why $$arg min_x ||Ax - b||_2 equiv arg min_x ||Ax - b||_2^2 $$?
Thank you!
linear-algebra regression
$endgroup$
I am reading online scribe notes on linear regression and I am being quite confused (since did not do linear algebra for a while) on how Pythagorean theorem is being applied here:
First of all I am confused of what "cost" is and then how we get from
$$arg min_x ||A(x-x') - bb'||_2^2 $$ to
$$arg min_x ||A(x-x')||_2^2 - ||b'||_2^2 $$ ? E.g I need some intuition and mathematical explanation of what is happening here and how Pythagorean theorem is used in vector space.
Also why $$arg min_x ||Ax - b||_2 equiv arg min_x ||Ax - b||_2^2 $$?
Thank you!
linear-algebra regression
linear-algebra regression
asked Dec 7 '18 at 19:54
YohanRothYohanRoth
6261714
6261714
add a comment |
add a comment |
1 Answer
1
active
oldest
votes
$begingroup$
The cost is $argmin_x ||Ax-b||_2^2$, and one aims at finding a minimizer $x$.
The "Pythagorean theorem" refers to the fact that
$$||x+y||_2^2 = ||x||_2^2 + ||y||_2^2$$ whenever $x perp y$.
In general (Hilbert spaces over $Bbb{C}$), we have
begin{align}
||Ax-b||_2^2 &= langle A(x-x') - b', A(x-x') - b' rangle \
&= langle A(x-x'), A(x-x')rangle - 2mathop{{rm Re}}(langle A(x-x'), b'rangle) + langle b',b'rangle \
&= ||A(x-x')||_2^2 + 0 + ||b'||_2^2 tag{$b' perp A(x-x')$} \
&= ||A(x-x')||_2^2 + ||b'||_2^2
end{align}
To address that last problem, observe that for any strictly increasing function $f$, $x < y iff f(x) < f(y)$ by defintion, so $f(p(x))< f(p(y)) iff p(x) < p(y)$. Take $f(x) = x^2$ on $[0,infty)$ to finish the argument.
$endgroup$
$begingroup$
I am not very clear on this math notation. What is x' or b'? What is Re? )
$endgroup$
– YohanRoth
Dec 7 '18 at 21:10
$begingroup$
I use $x'$ and $b'$ as in your notes. $Re$ means real part, but that's irrelevant to this question since you're dealing with real-valued data. I put this so that it's also true in Hilbert spaces.
$endgroup$
– GNUSupporter 8964民主女神 地下教會
Dec 7 '18 at 21:15
$begingroup$
one more question . Why b = Ax' + b' must be orthogonal to col space of A?
$endgroup$
– YohanRoth
Dec 7 '18 at 21:28
$begingroup$
@YohanRoth Sorry for late reply. I've made a typo. It should be $b'$ instead, as in the notes.
$endgroup$
– GNUSupporter 8964民主女神 地下教會
Dec 8 '18 at 9:11
add a comment |
Your Answer
StackExchange.ifUsing("editor", function () {
return StackExchange.using("mathjaxEditing", function () {
StackExchange.MarkdownEditor.creationCallbacks.add(function (editor, postfix) {
StackExchange.mathjaxEditing.prepareWmdForMathJax(editor, postfix, [["$", "$"], ["\\(","\\)"]]);
});
});
}, "mathjax-editing");
StackExchange.ready(function() {
var channelOptions = {
tags: "".split(" "),
id: "69"
};
initTagRenderer("".split(" "), "".split(" "), channelOptions);
StackExchange.using("externalEditor", function() {
// Have to fire editor after snippets, if snippets enabled
if (StackExchange.settings.snippets.snippetsEnabled) {
StackExchange.using("snippets", function() {
createEditor();
});
}
else {
createEditor();
}
});
function createEditor() {
StackExchange.prepareEditor({
heartbeatType: 'answer',
autoActivateHeartbeat: false,
convertImagesToLinks: true,
noModals: true,
showLowRepImageUploadWarning: true,
reputationToPostImages: 10,
bindNavPrevention: true,
postfix: "",
imageUploader: {
brandingHtml: "Powered by u003ca class="icon-imgur-white" href="https://imgur.com/"u003eu003c/au003e",
contentPolicyHtml: "User contributions licensed under u003ca href="https://creativecommons.org/licenses/by-sa/3.0/"u003ecc by-sa 3.0 with attribution requiredu003c/au003e u003ca href="https://stackoverflow.com/legal/content-policy"u003e(content policy)u003c/au003e",
allowUrls: true
},
noCode: true, onDemand: true,
discardSelector: ".discard-answer"
,immediatelyShowMarkdownHelp:true
});
}
});
Sign up or log in
StackExchange.ready(function () {
StackExchange.helpers.onClickDraftSave('#login-link');
});
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Post as a guest
Required, but never shown
StackExchange.ready(
function () {
StackExchange.openid.initPostLogin('.new-post-login', 'https%3a%2f%2fmath.stackexchange.com%2fquestions%2f3030307%2fpythagorean-theorem-in-linear-regression-using-matrix-notation%23new-answer', 'question_page');
}
);
Post as a guest
Required, but never shown
1 Answer
1
active
oldest
votes
1 Answer
1
active
oldest
votes
active
oldest
votes
active
oldest
votes
$begingroup$
The cost is $argmin_x ||Ax-b||_2^2$, and one aims at finding a minimizer $x$.
The "Pythagorean theorem" refers to the fact that
$$||x+y||_2^2 = ||x||_2^2 + ||y||_2^2$$ whenever $x perp y$.
In general (Hilbert spaces over $Bbb{C}$), we have
begin{align}
||Ax-b||_2^2 &= langle A(x-x') - b', A(x-x') - b' rangle \
&= langle A(x-x'), A(x-x')rangle - 2mathop{{rm Re}}(langle A(x-x'), b'rangle) + langle b',b'rangle \
&= ||A(x-x')||_2^2 + 0 + ||b'||_2^2 tag{$b' perp A(x-x')$} \
&= ||A(x-x')||_2^2 + ||b'||_2^2
end{align}
To address that last problem, observe that for any strictly increasing function $f$, $x < y iff f(x) < f(y)$ by defintion, so $f(p(x))< f(p(y)) iff p(x) < p(y)$. Take $f(x) = x^2$ on $[0,infty)$ to finish the argument.
$endgroup$
$begingroup$
I am not very clear on this math notation. What is x' or b'? What is Re? )
$endgroup$
– YohanRoth
Dec 7 '18 at 21:10
$begingroup$
I use $x'$ and $b'$ as in your notes. $Re$ means real part, but that's irrelevant to this question since you're dealing with real-valued data. I put this so that it's also true in Hilbert spaces.
$endgroup$
– GNUSupporter 8964民主女神 地下教會
Dec 7 '18 at 21:15
$begingroup$
one more question . Why b = Ax' + b' must be orthogonal to col space of A?
$endgroup$
– YohanRoth
Dec 7 '18 at 21:28
$begingroup$
@YohanRoth Sorry for late reply. I've made a typo. It should be $b'$ instead, as in the notes.
$endgroup$
– GNUSupporter 8964民主女神 地下教會
Dec 8 '18 at 9:11
add a comment |
$begingroup$
The cost is $argmin_x ||Ax-b||_2^2$, and one aims at finding a minimizer $x$.
The "Pythagorean theorem" refers to the fact that
$$||x+y||_2^2 = ||x||_2^2 + ||y||_2^2$$ whenever $x perp y$.
In general (Hilbert spaces over $Bbb{C}$), we have
begin{align}
||Ax-b||_2^2 &= langle A(x-x') - b', A(x-x') - b' rangle \
&= langle A(x-x'), A(x-x')rangle - 2mathop{{rm Re}}(langle A(x-x'), b'rangle) + langle b',b'rangle \
&= ||A(x-x')||_2^2 + 0 + ||b'||_2^2 tag{$b' perp A(x-x')$} \
&= ||A(x-x')||_2^2 + ||b'||_2^2
end{align}
To address that last problem, observe that for any strictly increasing function $f$, $x < y iff f(x) < f(y)$ by defintion, so $f(p(x))< f(p(y)) iff p(x) < p(y)$. Take $f(x) = x^2$ on $[0,infty)$ to finish the argument.
$endgroup$
$begingroup$
I am not very clear on this math notation. What is x' or b'? What is Re? )
$endgroup$
– YohanRoth
Dec 7 '18 at 21:10
$begingroup$
I use $x'$ and $b'$ as in your notes. $Re$ means real part, but that's irrelevant to this question since you're dealing with real-valued data. I put this so that it's also true in Hilbert spaces.
$endgroup$
– GNUSupporter 8964民主女神 地下教會
Dec 7 '18 at 21:15
$begingroup$
one more question . Why b = Ax' + b' must be orthogonal to col space of A?
$endgroup$
– YohanRoth
Dec 7 '18 at 21:28
$begingroup$
@YohanRoth Sorry for late reply. I've made a typo. It should be $b'$ instead, as in the notes.
$endgroup$
– GNUSupporter 8964民主女神 地下教會
Dec 8 '18 at 9:11
add a comment |
$begingroup$
The cost is $argmin_x ||Ax-b||_2^2$, and one aims at finding a minimizer $x$.
The "Pythagorean theorem" refers to the fact that
$$||x+y||_2^2 = ||x||_2^2 + ||y||_2^2$$ whenever $x perp y$.
In general (Hilbert spaces over $Bbb{C}$), we have
begin{align}
||Ax-b||_2^2 &= langle A(x-x') - b', A(x-x') - b' rangle \
&= langle A(x-x'), A(x-x')rangle - 2mathop{{rm Re}}(langle A(x-x'), b'rangle) + langle b',b'rangle \
&= ||A(x-x')||_2^2 + 0 + ||b'||_2^2 tag{$b' perp A(x-x')$} \
&= ||A(x-x')||_2^2 + ||b'||_2^2
end{align}
To address that last problem, observe that for any strictly increasing function $f$, $x < y iff f(x) < f(y)$ by defintion, so $f(p(x))< f(p(y)) iff p(x) < p(y)$. Take $f(x) = x^2$ on $[0,infty)$ to finish the argument.
$endgroup$
The cost is $argmin_x ||Ax-b||_2^2$, and one aims at finding a minimizer $x$.
The "Pythagorean theorem" refers to the fact that
$$||x+y||_2^2 = ||x||_2^2 + ||y||_2^2$$ whenever $x perp y$.
In general (Hilbert spaces over $Bbb{C}$), we have
begin{align}
||Ax-b||_2^2 &= langle A(x-x') - b', A(x-x') - b' rangle \
&= langle A(x-x'), A(x-x')rangle - 2mathop{{rm Re}}(langle A(x-x'), b'rangle) + langle b',b'rangle \
&= ||A(x-x')||_2^2 + 0 + ||b'||_2^2 tag{$b' perp A(x-x')$} \
&= ||A(x-x')||_2^2 + ||b'||_2^2
end{align}
To address that last problem, observe that for any strictly increasing function $f$, $x < y iff f(x) < f(y)$ by defintion, so $f(p(x))< f(p(y)) iff p(x) < p(y)$. Take $f(x) = x^2$ on $[0,infty)$ to finish the argument.
edited Dec 8 '18 at 9:09
answered Dec 7 '18 at 20:21
GNUSupporter 8964民主女神 地下教會GNUSupporter 8964民主女神 地下教會
12.8k72445
12.8k72445
$begingroup$
I am not very clear on this math notation. What is x' or b'? What is Re? )
$endgroup$
– YohanRoth
Dec 7 '18 at 21:10
$begingroup$
I use $x'$ and $b'$ as in your notes. $Re$ means real part, but that's irrelevant to this question since you're dealing with real-valued data. I put this so that it's also true in Hilbert spaces.
$endgroup$
– GNUSupporter 8964民主女神 地下教會
Dec 7 '18 at 21:15
$begingroup$
one more question . Why b = Ax' + b' must be orthogonal to col space of A?
$endgroup$
– YohanRoth
Dec 7 '18 at 21:28
$begingroup$
@YohanRoth Sorry for late reply. I've made a typo. It should be $b'$ instead, as in the notes.
$endgroup$
– GNUSupporter 8964民主女神 地下教會
Dec 8 '18 at 9:11
add a comment |
$begingroup$
I am not very clear on this math notation. What is x' or b'? What is Re? )
$endgroup$
– YohanRoth
Dec 7 '18 at 21:10
$begingroup$
I use $x'$ and $b'$ as in your notes. $Re$ means real part, but that's irrelevant to this question since you're dealing with real-valued data. I put this so that it's also true in Hilbert spaces.
$endgroup$
– GNUSupporter 8964民主女神 地下教會
Dec 7 '18 at 21:15
$begingroup$
one more question . Why b = Ax' + b' must be orthogonal to col space of A?
$endgroup$
– YohanRoth
Dec 7 '18 at 21:28
$begingroup$
@YohanRoth Sorry for late reply. I've made a typo. It should be $b'$ instead, as in the notes.
$endgroup$
– GNUSupporter 8964民主女神 地下教會
Dec 8 '18 at 9:11
$begingroup$
I am not very clear on this math notation. What is x' or b'? What is Re? )
$endgroup$
– YohanRoth
Dec 7 '18 at 21:10
$begingroup$
I am not very clear on this math notation. What is x' or b'? What is Re? )
$endgroup$
– YohanRoth
Dec 7 '18 at 21:10
$begingroup$
I use $x'$ and $b'$ as in your notes. $Re$ means real part, but that's irrelevant to this question since you're dealing with real-valued data. I put this so that it's also true in Hilbert spaces.
$endgroup$
– GNUSupporter 8964民主女神 地下教會
Dec 7 '18 at 21:15
$begingroup$
I use $x'$ and $b'$ as in your notes. $Re$ means real part, but that's irrelevant to this question since you're dealing with real-valued data. I put this so that it's also true in Hilbert spaces.
$endgroup$
– GNUSupporter 8964民主女神 地下教會
Dec 7 '18 at 21:15
$begingroup$
one more question . Why b = Ax' + b' must be orthogonal to col space of A?
$endgroup$
– YohanRoth
Dec 7 '18 at 21:28
$begingroup$
one more question . Why b = Ax' + b' must be orthogonal to col space of A?
$endgroup$
– YohanRoth
Dec 7 '18 at 21:28
$begingroup$
@YohanRoth Sorry for late reply. I've made a typo. It should be $b'$ instead, as in the notes.
$endgroup$
– GNUSupporter 8964民主女神 地下教會
Dec 8 '18 at 9:11
$begingroup$
@YohanRoth Sorry for late reply. I've made a typo. It should be $b'$ instead, as in the notes.
$endgroup$
– GNUSupporter 8964民主女神 地下教會
Dec 8 '18 at 9:11
add a comment |
Thanks for contributing an answer to Mathematics Stack Exchange!
- Please be sure to answer the question. Provide details and share your research!
But avoid …
- Asking for help, clarification, or responding to other answers.
- Making statements based on opinion; back them up with references or personal experience.
Use MathJax to format equations. MathJax reference.
To learn more, see our tips on writing great answers.
Sign up or log in
StackExchange.ready(function () {
StackExchange.helpers.onClickDraftSave('#login-link');
});
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Post as a guest
Required, but never shown
StackExchange.ready(
function () {
StackExchange.openid.initPostLogin('.new-post-login', 'https%3a%2f%2fmath.stackexchange.com%2fquestions%2f3030307%2fpythagorean-theorem-in-linear-regression-using-matrix-notation%23new-answer', 'question_page');
}
);
Post as a guest
Required, but never shown
Sign up or log in
StackExchange.ready(function () {
StackExchange.helpers.onClickDraftSave('#login-link');
});
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Post as a guest
Required, but never shown
Sign up or log in
StackExchange.ready(function () {
StackExchange.helpers.onClickDraftSave('#login-link');
});
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Post as a guest
Required, but never shown
Sign up or log in
StackExchange.ready(function () {
StackExchange.helpers.onClickDraftSave('#login-link');
});
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Post as a guest
Required, but never shown
Required, but never shown
Required, but never shown
Required, but never shown
Required, but never shown
Required, but never shown
Required, but never shown
Required, but never shown
Required, but never shown