#linear-algebra

2 messages Β· Page 139 of 1

gray dust
#

we have that if T is linear then T(0)=0. what's the contrapositive of this statement?

cobalt vale
#

I got it. thanks man @gray dust

#

although, i have another question

#

do you need to be in row echelon to identify pivots ?

#

for example

#

-3 5 -1

#

0 2 1

#

0 0 -3

#

we can clearly see that the pivot positions are -3 2 and 3 right ? or would u need to row reduce it to prove it

half ice
#

You're correct that's enough

#

Upper triangular will reduce to identity

#

If nothing on the diagonal is 0

gray dust
#

@cobalt vale it's already in ref so you can immediately count off pivots

gritty frigate
#

If a line in R3 has a director vector with two components = 0. The line will have infinite projecting planes?

half ice
#

Director vector lol

gritty frigate
#

I dont know the translation hahaha

#

Sorry

half ice
#

Np it's just funny. Unfortunately I don't follow the rest. Projecting planes?

gritty frigate
#

Yeah, It is a plane that has the line and is perpedicular to any xy,yz or zx plane

#

Again, sorry if it is not the name of the concept, translating math is kinda hard

half ice
#

The line will be parallel to the x-axis, or the y-axis, or the z-axis

gritty frigate
#

Yep

half ice
#

Yeah and will be perpendicular to either of those planes

gritty frigate
#

And the plane, which that line is normal to will have infinite "projecting planes" ?

fair vector
#

Can anyone explain to me the concept of spans ? Like what does it mean when something spans R^n

native rampart
#

Ok, Let's say you have a set of vectors

#

And you consider the set of all linear combinations of said vectors,that set is called the span of your vectors{say you chose (1,0,0) and (0,1,0) in vector space R^3 ,then all vectors of form x(1,0,0)+y(0,1,0) will be the span of {(1,0,0),(0,1,0)}}

#

@fair vector

#

Now if the span of vectors is R^n ,you say the vectors span R^n

fair vector
#

Yea got it i understood this but if there is a specific case for example in exams they give random cases lk {3,1,2} and {4,4,2} does this span R^3 how should i proceed

#

is their any tricks

native rampart
#

You know what a basis is?

fair vector
#

yea

native rampart
#

Check the number of linearly independent vectors in your spanning set

#

If it is equal to dim(space),it spans. Otherwise it doesn't span

fair vector
#

oh okay thanks i didnt know all this

robust pond
#

i have a dumb question

dusky epoch
#

shoot

robust pond
#

this arrowed part

#

i know ive done this before but now i cannot for the life of me remember

#

you find the free variables

#

then you set it to be some parameter

#

im not understanding how to use this though

#

this is for finding the bases of the null space

#

most links i find skip over the details

#

or maybe the details are included and im not seeing it

dusky epoch
#

you can switch back to the equation view

#

you have
x1 + x4 = 0
x2 + x4 = 0
x3 + x4 = 0

robust pond
#

oh you just split it into a system

#

i figured that was not valid for a second

#

im not sure why

#

i got it πŸ™‡β€β™‚οΈ thanks

lucid cedar
#

im trying to understand a proof that i am working through that contains this line

#

It seems rather non obvious that this statement is true and it is given without explanation

native rampart
#

Is your doubt about T being linear?

lucid cedar
#

no

dusky epoch
#

it's a geometric series

lucid cedar
#

oh

#

wait

#

ann your a jeniuos

#

thank you!

opaque plover
#

Hello, I am kinda stuck with this problem: How many equivalence class are there for this relation?

marble lance
#

@opaque plover do you still need help?

opaque plover
#

@marble lance yes please

marble lance
#

Sorry, will ping you in 20 min

opaque plover
#

alright thanks

marble lance
#

@opaque plover nvm

#

Do you know that for every integer a and positive integer b, you can find integers q and r so that a = bq + r and 0 <= r < b?

opaque plover
#

How do you come to this conclusion, because I cant follow you, sry

dusky epoch
#

this is just division with remainder

#

for example, if a = 258 and b = 13, then you get 258 = 13*19 + 11

opaque plover
#

yeah i got that

marble lance
#

Okay, great

#

So now you have to show two things: 0, 1,2,...,n-1 belongs to different equivalence classes; and every integer belongs to one of those classes

#

That would mean there are n equivalence classes: one for each possible remainder after division by n

#

Does that make sense? @opaque plover

#

Consider 0 <= i < j < n-1. Then 0 < j-i < n, so j-i is not a multiple of n. So i and j belong to distinct equivalence classes. This proves the first part

#

Now consider any integer a. Then a = nq + r, so a-r = nq. Then a belongs to the same equivalence class as its remainder r. This proves the second part.

opaque plover
#

ohhh yeah that makes sense, how can you solve this so fast lol

marble lance
#

I'm already familiar with the proof. This is a very well known equivalence relation

opaque plover
#

alright thanks you very much

#

and you too @dusky epoch

marble lance
#

Np

calm hamlet
#

If I want to prove the vector space of real sequences with limit=0 H is a hyperplane of the vector space of real convergent sequences N, is it a valid proof to say:

I take any real convergent sequence (a_n) of limit L, I build the sequence (b_n) such as b_n=A_n -L belongs to H, and then each sequence from N can be separated in one sequence from H and one constant sequence, then H is a hyperplane of N?

dusky epoch
#

what's your defn of a hyperplane

magic light
#

https://i.imgur.com/6hmO665.png
Hello, I need an answer here if possible

I am asked if there exists a linear transformation
T, s.t Ker(T)=V AND
3 is a eigenvalue of T

Now, for Ker(T)=V that means it spans all the same
but 3 is an eigenvector meaning
T(v) = 3v for some v

because v != 0, then 3v != 0 and thus T(v) is NOT in Ker(T), does that disprove the existence of such a transformation?
Because on one hand KerT = V
on the other hand here's a v that is only in the Image ?

#

V, T, M all written in the picture

wintry steppe
native rampart
#

There is no additive identity

wintry steppe
#

Ohh

#

Thanks lol

magic light
#

<@&286206848099549185> question above

#

<@&286206848099549185> pls, it's semi-urgent

idle current
#

<@&286206848099549185> pls, it's semi-urgent
@magic light Why is that?

wintry steppe
#

Can the coefficients on linear combinations be irrational?

mortal stream
#

why not

wintry steppe
#

Well I dont see how anything would be linearly independent

native rampart
#

Well I dont see how anything would be linearly independent
@wintry steppe they can be anything from any field

#

(Ok,The field your vector space is defined with)

mortal stream
#

for your vector space question, consider two polynomials, one with leading coefficient 1 and one with -1

#

add them and you'll see that the result is not in the set

#

hence it's not closed

wintry steppe
#

So wait what would an example of a linearly independent system in R be then?

magic light
#

@idle current because I submitted it as a test answer, and I got partial points, wanted to know if I can get a little more depending on how wrong I was

#

there's only a limited amount of days that I can ask for a review

wintry steppe
#

The reason I ask is because im struggling to see how any system could be independent in R if we can just take any real number as coefficients

slate haven
#

should this be evaluated as $(x_1 \wedge x_3) - (2x_1 \wedge x_2) + (4x_2 \wedge x_3)$ or as $x_1 \wedge (x_3 - 2x_1) \wedge (x_2 + 4x_2) \wedge x_3$
I'm slightly confused as i can't find anything in the notes. My feeling tells me that the first evaluation should be correct since it's kind of a multiplication but idk, might be a dumb question

stoic pythonBOT
slate haven
#

ok well since the exterior product fulfills distributivity it should definitely be the first evaluation order, right ?

scarlet trail
#

2t+8=3

stuck stratus
dawn remnant
#

Have you solved the characteristic equation?

stuck stratus
#

yeah

#

lambda = 0 or lambda = k+2

#

am(0) = 2, am(k+2) = 1

quaint marlin
#

@dusky epoch can you help me with this please?

dusky epoch
#

why am i of all people getting pung out of the blue

dawn remnant
#

Oh, they pinged me first, then deleted the message πŸ˜›

quaint marlin
#

Sorry 😭😭

#

Well can one of you guys help me please lol

dusky epoch
#

k so what are you having trouble with

quaint marlin
#

Like where do I start?

#

Should I take the determinant?

dusky epoch
#

if you want

quaint marlin
#

And then solve for lambda?

dusky epoch
#

before you solve for lambda you need an equation to solve for lambda

quaint marlin
#

Could I use the 4x-y+lambda z = 8?

dusky epoch
#

depending on what you mean by "use"

#

do you want me to pretend this question admits one and only one solution path & tell you exactly what steps go into it?

quaint marlin
#

Yes please

#

I only need the steps and can go from there

dusky epoch
#

evaluate determinant, set determinant equal to zero, solve for lambda

quaint marlin
#

Ok, that's what I've done hahah

#

I was just making sure that was correct

#

Tyvm

cunning arch
#

wouldn't it just be the same?

wintry steppe
#

on top its the solution for Ax=b, but they ask for the solution for Ax=0

#

If you need an hint: || Suppose Ax = b and Ay= b. Then what is A(x-y)?||

cunning arch
#

Sorry for the really late response, I'm still a bit confused. A(x-y) would just be b-b=0 (?)

#

Are we supposed to solve the system 5+t+3r=0, 1+t=0,-2+5r=0,3+t+r=0? Because that ends up inconsistent so no solution set, so I think that's incorrect

wintry steppe
#

no. you are not given A

#

what they tell you is that all vectors of that form where t and r are numbers are solution for the system Ax=b

#

and they want you to find the solution to Ax=0 from just that information. They do not tell you what A and b are.

lucid cedar
#

Im trying to solve for the eigenvalues and vectors of the Map:
T(w,z) = (-z,w)

#

Axlers explanation for finding the Evalues is lengthy and easy to follow. They are i && -i. His explanation of the vectors is just "Therefore the vectors are (w,wi),(w,-wi)" and i just cant see how he came to this conclusion.

#

nevermind i see how it works

wintry steppe
#

how did i mess up on the gaussian elimination to incorrectly determine that this matrix's vectors are linearly independent

brisk fractal
#

(1,4) + (-1,4) = (0,8)

#

also, avoid writing that the matrices are equal, because they're not

#

they're equivalent via elementary matrices

wintry steppe
#

ok thanks

brisk fractal
#

I'm also not really sure what he wants the reader to say about their relation

wintry steppe
#

πŸ€”

#

can someone confirm that rankA = rankA^T = nullA = nullA^T = 1 for this matri
Sounds right

#

relation
It's the same space (1, i)?

#

Maybe there's a term for that or something

brisk fractal
#

I thought that was interesting, I noticed that A^2 = 0 and figured there was something to it

#

there's probably some wacky spectral theory shit you can do with nilpotent matrices

cobalt vale
#

can a set of vectors be a basis

#

if they have no solution

brisk fractal
#

wdym by "have no solution"

cobalt vale
#

the system has no solution

brisk fractal
#

oh, like if Ax does not equal b for some b in the codomain

cobalt vale
#

that system has no solution

#

so its not linearly independent

#

therefore it's not a basis of r2

#

or am i missing something here

#

ax=0

#

oh wait yikes

#

i performed elementary operations wrong

brisk fractal
#

if you put it into a matrix, you can reduce it to
0 2
20 0
which implies that nullity of the matrix is 0

cobalt vale
#

x1 and x2 =0

#

sporry im dumb

#

sorry sorry

brisk fractal
#

no worries

cobalt vale
#

can someone double check my answers ?

#

true false true true false false

brisk fractal
#

d is false, that's only true for surjective linear maps

#

everything else looks good

cobalt vale
#

i dont agree with d

brisk fractal
#

better explanation for d: consider when rank(A) < dim(V)

#

then colA doesn't span W

cobalt vale
#

i was thinking more like. for any vector w in W there exists at least one v in V such that T(v)= w=>T(V) = W

#

that was my explanation for it

brisk fractal
#

see, that's not necessarily true though

#

let T be a matrix, if there isn't a pivot in every row, then that's not true

cobalt vale
#

OHHH

#

man i was thinking too deep lol

brisk fractal
#

the pivot stuff is confusing, but once you think with rank stuff it becomes 30x easier

cobalt vale
#

i agree!

#

do you mind one last answer check ?

brisk fractal
#

hit me

cobalt vale
#

true false false false true false true true true true

#

im guessing c aint right ?

brisk fractal
#

true, false, false, false, true, true, false, false, false, false I think

cobalt vale
#

isnt F true ?

#

well actually, im guessing free variables

brisk fractal
#

I wrote f is true

cobalt vale
#

no f is false

brisk fractal
#

I'm not 100%, I just finished the exact same material

cobalt vale
#

a pivot in the last column would make it inconsistent

brisk fractal
#

pivot in each row is consistency, pivot in each column is spanning

cobalt vale
#

because you get 1 = 0

brisk fractal
#

it's rows not columns in F

cobalt vale
#

right

#

lets say

#

100

#

001

#

this is incosistent

#

because 1=0

brisk fractal
#

sorry, pivot in each row of the coefficient matrix

#

not the augmented matrix

#

I guess if they're specifically referring the the augmented, then no, but none of the pivot rules refer to augmented matrices iirc

wintry steppe
#

How to debunk every single mathematical theory:

this is incosistent
because 1=0

cobalt vale
#

also i think the last 4 are def true

#

due to theorems

brisk fractal
#

Sorry, yeah, i didn't read the "augmented" part because that fucks everything up

#

but h is false, that would imply that columns have pivots, not rows

#

i is true, I was dumb

#

j is equivalent to h I believe

#

wait fuck yeah I'm being dumb

#

j is true my bad

dense sonnet
#

Can anyone recomend a book that talks about the history of the hessian matrix? On wiki there really isn't any history behind it. Thank you

cunning arch
#

ok i'm legitimately stupid. despite the help earlier, i still don't know what to do with this. i understand what it's asking, just not how to find it (?). could someone guide me through this one please?

half ice
#

Just for ease of writing, I'll label those vectors such that:
x = u + tv + rw

It is the case that:
Au = b
Just by setting t = r = 0

Then,
Ax = b
A(u + tv + rw) = b
Au + A(tv + rw) = b
b + A(tv + rw) = b
A(tv + rw) = 0

cunning arch
#

oooooooh

#

wow that ended up being a lot more simple than i thought. thank you so much @half ice, that made a lot more sense :)

half ice
#

Np! Feel free to ask if you need anything else

wintry steppe
#

Need help

#

Urgently

limber sierra
#

why urgently?

soft burrow
#

the "(1 point)" might be a hint lmao

#
  1. Requesting help during an exam is a bannable offense.
    #rules
cunning arch
#

so uh, am i the only one they cold-DMed to get help? nice blurryeyes

limber sierra
#

i'd normally rather give the benefit of the doubt, but considering the cold DM...

#

hrm...

wintry steppe
#

It's not really urgent

#

Not an exam

#

Still here

#

Just urgent because I intend on sleeping soon

#

Anyways I figured it out

elfin mist
#

what's a good place to learn about dual and double dual spaces?

native rampart
#

Hoffman kunze had a good introduction,I think

elfin mist
#

alright cool I'll check it out

tired garden
dusky epoch
#

does x = [0;0;0] satisfy the equation a Β· x + 2 = 0 or not

tired garden
#

Obviously not since we add 2 to it after

#

Is that it?

#

any vector multiplied by 0 vector will give 0s

#

Adding 2 will give us 2

dusky epoch
#

yes that is it

tired garden
#

Thanks

#

I was overthinking it

scenic vapor
#

given a semi orthogonal matrix U

#

and:

#

how can i show that it is invertible

unique yew
#

The question asks to find the set of solutions for x+2y+3z=0. How can this be done?

scenic vapor
#

is a "square semi orthogonal matrix" always going to be orthogonal?

#

i thought a square matrix could not be semi orthogonal

quasi vale
#

@unique yew We can solve for x or y or z if we have values for two variables. So put y=t, where t is any real number, z=s, where s is any real number and solve for x.

unique yew
#

@quasi vale how is this going to prove the number of solutions?

#

also thanks for answering

whole nimbus
#

can anyone help me out on the 3rd step? the solution is so vague in microsoft's math solver.

wintry sphinx
#

@dim venture it means that if U is a subset of W, then Uperp (orthogonal complement) a subset of Wperp

honest imp
brisk fractal
#

is this supposed to be a "gotcha!"?

#

I did det(A^T) = (-1)^n*det(A) = det(A), and if n is odd then -det(A) = det(A) and so det(A) = 0

#

if n is even then it's not necessarily true right

pallid rampart
#

Let n be odd, then look at the $n\times n$ matrix $(a_{i,j})$ with $a_{i,n-i+1}=\begin{cases}1&i=1,2,...,n/2\-1&i>n/2\end{cases}$ and $0$ everywhere else

stoic pythonBOT
pallid rampart
#

Basically the matrix $\begin{bmatrix}0&0&\dots&0&0&\dots&0&1\0&0&\dots&0&0&\dots&1&0\\vdots&\vdots&\ddots&\vdots&\vdots&&\vdots&\vdots\0&0&\dots&0&1&\dots&0&0\0&0&\dots&-1&0&\dots&0&0\\vdots&\vdots&&\vdots&\vdots&\ddots&\vdots&\vdots\0&-1&\dots&0&0&\dots&0&0\-1&0&\dots&0&0&\dots&0&0\end{bmatrix}$

#

Let's see if texit can render this abomination

stoic pythonBOT
true talon
#

Ive made the system into an augmented matrix. Then tried to make it into row echelon. After this I tried solving for the different solutions

#

I’m unsure if my equations and approach are correct

#

I’m pretty sure my iii is incorrect

pseudo abyss
#

So geometrically these are two lines through origin...they have a unique solution i.e $(0,0)$ if and only if their slopes are different i.e $12\neq (1-\lambda)(2-\lambda)$...so infinitely many solutions are possible when $12= (1-\lambda)(2-\lambda)$ which i am sure you can solve...

stoic pythonBOT
true talon
#

@pseudo abyss thanks! could you explain how you equated both the equations and figured that out tho?

pseudo abyss
#

Just equate the slopes of the two lines...first one is $(\lambda-1)/12$...second one is $1/(\lambda -2)$

stoic pythonBOT
true talon
#

yeah i get that, but why is that when they are equal to 12 that they have infinite solutions, but unique solutions when they arent equal to 12

wise flare
#

hi

#

anybody here ?

dusky epoch
#

no obviously in a server of thousands nobody is ever online and this server has been empty and desolate for years /s

#

@wise flare is there a question you wanted to ask

pure thistle
#

Can someone tell me how the matrix (f:e,e) and (f:a,a) are calculated? I need the methodology. No need to explain to me just tell me what to google

half ice
#

@pure thistle
This looks like the outputs of f evaluated at e1, e2, then arranged into a matrix. Then the other matrix is the outputs of f evaluated at a1, a2

#

I wouldn't know what to tell you to Google, I don't know why they're doing this

#

Have some context, perhaps?

pure thistle
#

The subject is linear algebra 2. This one is the first thing written in notes and i dont remember if it was in linear algebra 1 or I'll learn it in linear algebra 2

half ice
#

It could be a linear operator represented in two different basis?

pure thistle
#

Yeah probably something like that

restive forge
#

someone want to tell me how to find the image of transformations

robust pond
#

if someone could help me understand why my answer on this problem is wrong;

#

here's the givens I have

#

I'm trying to find [id]_C ^B

#

I'm attempting to use:

#

so i figure that

#

$[\text{id}]_c ^b = \begin{pmatrix} 1 & -3 \ -1 & -3 \end{pmatrix} \begin{pmatrix} \sfrac{1}{3} & \sfrac{1}{9} \ \sfrac{-1}{3} & \sfrac{2}{9} \end{pmatrix} = \begin{pmatrix} \sfrac{4}{3} & \sfrac{-5}{9} \ \sfrac{2}{3} & \sfrac{-7}{9} \end{pmatrix}$

stoic pythonBOT
robust pond
#

this is wrong though, im struggling to see why

half forge
#

can someone explain too me why this is false

robust pond
#

oh i thought you were supporting me

#

and not just stealing the channel πŸ”ͺ

golden drum
#

@half forge If V and W have the same dimension, you can map a basis from V to a basis from W, that's a Isomorphism

native rampart
#

You don't speak of dimensions if V and W are infinite dimensional

half forge
#

so then is it false or true

dusky epoch
#

2 finite dim vector spaces over a field F

not specified to be findim & over the same field

half forge
#

im lost loll

#

is this false ?

gloomy spoke
#

Why's the coefficient 1, other than (-1)^n?

#

Must not have been |c|(-1)^n where c is a constant?

gloomy spoke
#

Oh that seems too heavy to assume implicitly πŸ˜…

#

But even if that's the case, there doesn't seem done obvious reason why the equation will turn out to be as such... Can you guide me with a hint?

native rampart
#

I mean,The easiest thing to do is probably convert the matrix to a upper triangular matrix similar to it

#

The det(A) being product of eigenvalues and tr(A) being sum both follow immediately

gloomy spoke
#

What I understand without Leibniz is - the eighth values must form linear factors in the char polynomial, and the method of taking det will give that -1^n

#

Okay @native rampart, that seems quick. I'll think and revert if that's fine..

tacit sonnet
#

A Hilbert Space is just a linear vector space with some additional properties? Is my understanding correct?

gloomy spoke
#

Yes

tacit sonnet
#

Okay, thanks but I still am not able to understand it fully what it means... I looked it up on Google but the answers there just confused me... so,, can someone please explain it to me... just an overview would do

wintry steppe
brittle juniper
#

it's written in your picture

wintry steppe
#

oh

bold python
#

Assume that $T:V\to V$ is a linear transformation and that $\beta$ is a ordered basis for $V$ let $[T]\beta$ be the coordinate matrix to to the matrix $T$ with respect to $\beta$ then $$T \quad\text{is isomorphic}\quad \iff [T]\beta \quad\text{is invertible}$$ then if $T$ is isomorphic we have $[T^{-1}]\beta=([T]\beta)^{-1}$

stoic pythonBOT
bold python
#

how is $[T^{-1}]\beta = ([T]\beta)^{-1}$, why is this true?

stoic pythonBOT
native rampart
#

How did you write [T] post B?

bold python
#

T post B?

native rampart
#

(What do you call the small beta on the bottom)

bold python
#

ordered basis of V

native rampart
#

$T \mapsto [T]\beta$ so
$T^{-1} \mapsto [T^{-1}]
\beta$ implying
$TT^{-1} \mapsto [T]\beta [T^{-1}]\beta$
i.e.,
$[T]\beta[T^{-1}]\beta=I$

stoic pythonBOT
native rampart
#

Implying your result

#

@bold python

bold python
#

oh

#

thanks drunken

#

but

native rampart
#

Do you understand the isomorphism condition?

bold python
#

what justifies that you can write $[T]^{-1}\beta = ([T]^{-1}_\beta)$

stoic pythonBOT
bold python
#

yeah, that it holds the same structure under a transformation

native rampart
#

You know $[T]\beta [T^{-1}]\beta$=I

stoic pythonBOT
native rampart
#

Which means $[T^{-1}]\beta$ = ${[T]\beta}^{-1}$

stoic pythonBOT
bold python
#

OH

#

oh wow now i see it

#

thanks drunken

rotund jetty
#

If i have a linear transformation from a vector space to itself, do I need to write the input space and the output space wrt the same basis? For the purposes of representing it as a matrix

native rampart
#

You could just say it's an operator on V ,or whatever your vector space is

rotund jetty
#

oh yeah that is how the question is phrased

#

I have a linear operator on V, and need to show that I can get a matrix representation of some form or some other form

#

and I'm wondering if I can let the "input basis" and the "output basis" be different

native rampart
#

You could,but you generally wouldn't

rotund jetty
#

Ok, thanks!

waxen jacinth
#

Anyone that can explain this to me? The span is smth I just cant wrap my head around

#

No matter how many times I read the textbook, I don't get wts going on

robust pond
#

I'm on this topic rn too so kinda blind leading the blind but

#

at least on parts a and b

#

i think the dimension of the space is going to be the number of linearly independent columns you have

native rampart
#

Yes

robust pond
#

so to find b, you just need to find r and s to make those two columns linearly dependent

#

i dont think you can get to dim W = 1

#

just looking at the first and last rows

#

but thats my guess

native rampart
#

Yeah

robust pond
#

@waxen jacinth you prolly got recommended this already but the 3b1b video on span is good

#

once you can visualize or grok or understand span the rest sorta falls into place

#

at least thats how it was for me

waxen jacinth
#

Mhm

#

Thats what I was watching rn lol

#

Again

#

Kekw

robust pond
#

me too last night

#

cept the one on bases

#

i dont really like his videos on lin alg but i keep going back to them for some reason

rotund jetty
#

Do you know the definition of "span" diemia?

waxen jacinth
#

Well for 2 of those vectors, to be linearly indep, theyd need to fulfill these conditions:
r-2=1
r+2=/=1
s-1=0

#

I think

rotund jetty
#

not quite

waxen jacinth
#

A span is (probably gonna butcher the statement) a representation of the linear combinations of vectors

rotund jetty
#

well, sort of; the span is the set of all possible linear combinations

#

so it's every vector you can get to through linear combinations

waxen jacinth
#

Yeah, for example in 2D, u can get either all 2D vectors or just a line restricted by the vectors whose combinations u find

rotund jetty
#

So yeah for the dimension of the span of those vectors to be 3, all of them need to be linearly independent, so you need

$$a_1\begin{pmatrix}1 \ 0 \ 0 \ 0 \end{pmatrix} + a_2\begin{pmatrix}0 \ r - 2 \ s - 1 \ 0 \end{pmatrix} + a_3\begin{pmatrix}0 \ 2 \ r + 2 \ -3 \end{pmatrix} = \begin{pmatrix}0 \ 0 \ 0 \ 0 \end{pmatrix}$$

To have no solutions other than $a_1 = a_2 = a_3 = 0$

stoic pythonBOT
rotund jetty
#

if all three are linearly independent, they'll span something 3 dimensional

waxen jacinth
#

Yeah

#

It'll be at most an "infinite cube" I guess

#

In a 4D space

#

One plane remains untouched

#

In the 4D coordinate system

#

The other 3 are all filled in

rotund jetty
#

so you can literally just solve for the values of $r$ and $s$ that cause that equation to only have the solution where $a_1 = a_2 = a_3 = 0$

stoic pythonBOT
waxen jacinth
#

Can't I just plug them all in one matrix and make it so there are only 2 pivots?

#

Since the problem needs dim(W)=2

#

Oh nvm

#

Lul Im just tired excuse me ur correct

#

I havent had anything but 2 eggs today xd

#

Yes, so 3 pivots then

rotund jetty
#

wait weren't you solving dim(W) = 3?

waxen jacinth
#

I was, sorry I got confused

#

Ur correct

#

I mean both of them are there

rotund jetty
#

yes, but do dim(W) = 3 first

waxen jacinth
#

Both those questions should be solved, but the first one is dim(W)=3

rotund jetty
#

it's more direct

waxen jacinth
#

Yeah sorry xd

rotund jetty
#

just find the values of r and s so that the only solution to the linear dependence I wrote above is a_1 = a_2 = a_3 = 0

waxen jacinth
#

So, for that, I'd need
r-2=1
s-1=0
r+2 can be anything?

#

No, r+2 should be 0

#

As the last column will serve as the pivot column

#

On -3

#

I think.

#

Im confused

#

I mean, no matter what I do to these vectors, theyre never gonna be linearly independent cuz theres too many variables

rotund jetty
#

yes they very much are able to be linearly independent

#

I would suggest breaking it up row by row. First, look at the top row, then look at the bottom row. You can get rid of 2 of the variables right away

waxen jacinth
#

Maybe if the last vector is redundant

#

I suppose.

rotund jetty
#

So the first row gives you $a_1 + 0 + 0 = 0$, so $a_1 = 0$

stoic pythonBOT
rotund jetty
#

the last row gives you $0 + 0 - 3a_3 = 0$, so $a_3 = 0$

stoic pythonBOT
rotund jetty
#

so you now need to find the values of $r$ and $s$ that FORCE $a_2 = 0$ and you're done

stoic pythonBOT
rotund jetty
#

that gives you 3 linearly independent vectors, so a 3-dimensional subspace

waxen jacinth
#

So I guess

#

r=2, s=1?

#

Meh, idk

brittle juniper
#

There are a lot of projections that aren't injective pandaThink

normal falcon
#

basic question

#

(AB)^(-1) = B^-1 A^-1

#

right?

robust pond
rotund jetty
#

@waxen jacinth if r = 2 and s = 1, then a_2 = 5 or a_2 = 3 or a_2 = any real number works

#

you want either r or s to NOT equal 2

#

so that a_2 = 0 works

robust pond
#

@normal falcon yes

normal falcon
#

then why isn't the projection matrix the identity matrix?

#

P = A(A^T A)^-1 A^T

native rampart
#

How do I show that the projection is an injective linear transformation?
Show Ea=0 implies a=0

#

Usually it's not

hard coral
#

well and prove it is indeed linear. I'm pretty sure that does not necessarily imply linearity on its own

#

(take f(x) = e^x -1 for example, def. not linear as f(a)+f(b) = e^a + e^b - 2 β‰  e^(a+b) - 1 = f(a+b); but f(0) = e^0 - 1 = 1-1 = 0)

brisk fractal
#

Sorry, I posted this earlier in a question channel and I didn't get any help so I'll post again

#

Is this as straightforward as the matrix just being rows of Av_k, and so $A = (\lambda v_1, ..., \lambda v_n, Av_{n+1}, ..., Av_k)$, which is just the block triangular matrix since $\lambda v_n = \begin{pmatrix} 0 \ \vdots \ \lambda \ \vdots \ 0 \end{pmatrix} $ in that basis

stoic pythonBOT
wintry steppe
#

@brisk fractal seems like some spectral stuff

slow scroll
#

(To bacano) Yea I think you have the right idea. For a proof, u would just have to show that the action of the block matrix on basis vectors is the same as the action of A on basis vectors

stoic pythonBOT
wintry steppe
#

^ not super sure what to do

brisk fractal
#

(To bacano) Yea I think you have the right idea. For a proof, u would just have to show that the action of the block matrix on basis vectors is the same as the action of A on basis vectors
@slow scroll doesn't this just imply that the linear operators are equal, not necessarily the matrices? or does that not matter if they're in the same basis

calm hamlet
#

the image of the first k vectors v_i by A is Ξ»v_i, so you can write the first k rows of the matrix as the beginning of Ξ»β€’I_n.
For the other vectors, you have no information other than their image exists.
So you can write the first k rows as Ξ»β€’I_k on top and 0_kx(n-k) at the bottom, and the (n-k) other rows as a combination of some (n-k)xk matrix on top and some (n-k)x(n-k) matrix at the bottom since you have no information

hollow finch
#

suppose i have some m-dimensional subspace W of an n-dimensional vector space V with basis vectors w1,...,wm. if i put the coordinate vectors of each vector wi with respect to the standard basis of R^n as the rows of a matrix A, then the null space of A will contain every vector x in R^n for which the standard Euclidean inner product on R^n of the coordinate vector of wi and x is zero.
could i then say that a basis for the null space of A, v1,...,v(n-m), consists of coordinate vectors for a basis for the orthogonal complement of W back in V? or would that only work for a specific inner product on V?

slow scroll
#

that is just a consequence of working in the usual inner product on R^n (or C^n). You can't do this with arbitrary inner products

hollow finch
#

so you cant always use that null space method to find a basis for the orthogonal complement because the inner product might be different?

#

however if we're dealing with some definition of an inner product which results in exactly the same thing as the dot product of the coordinate vectors (a standard inner product i guess?), then it does still work though?

brisk fractal
#

okay well I've proven the two lemmas

#

I have no fucking clue of how to go about 1.9

#

the geometric multiplicity of an eigenvalue is just null(A-aI) right

#

and I assume they want you to fuck around with A so it can be arranged like in 1.8 so you can take the determinant using 1.7 to obtain some result about algebraic multiplicity of the same eigenvalue

#

but, for our space, we just know that Ker(A-aI) has a basis, not that that basis corresponds to the same basis at V

#

or at least to some finite subset

tulip basalt
#

is this a valid elementary row operation: (R3-1/3(R1+R2))?

brisk fractal
#

yes

tulip basalt
#

Thank you!

proper valley
#

Can anyone explain this

nocturne jewel
#

Is there an application for transposes? (context my course just introduced the transpose this week but I dont see the point in it)

dawn remnant
#

they will be useful in a lot of ways, yes.

proper valley
#

Can anyone help me with the problem?

#

I dont mind voice/zoom

dawn remnant
#

for one, there matrices for which $M^{-1} = M^T$ (orthogonal matrices, the O group), and also those for which $\bar{M}= M^T$ (where the bar is complex conjugate) (hermitian matrices), and both of these groups happen to be common in physics. @nocturne jewel

stoic pythonBOT
nocturne jewel
#

Yeah when M = M^T it's symmetric

dawn remnant
#

(orhogonal matrices conserve distances and so are basically rotations, whereas any transform from a basis to another happens to be a hermitian matrix, and a lot of physics uses coordinate transforms)

#

so yeah, transposes will be important.

nocturne jewel
#

Ok cause first glance is: Swap rows and columns βœ…

brisk fractal
#

if you have a linear combination of the vectors in the set, then one vector in the set is in the span of the other vectors and by definition the set is then linearly dependent

dreamy iron
#

yes.

brisk fractal
#

so then what are you proving exactly

#

by definition the set would be linearly dependent

dreamy iron
#

well, i don't think that it's by definition.

#

if you have a linear combination of the vectors in the set, then one vector in the set is in the span of the other vectors

I just proved this statement, no?

#

that's what i did.

#

adding a linear combination of the vectors in your set, will not let you leave your set

brisk fractal
#

the definition of linear dependence of a set of vectors is if $0 = \sum \alpha_n \mathbf{v}_n$ for nontrivial $\alpha_n$, i.e. $\mathbf{v}n = \frac{1}{\alpha_n} \sum{n \neq i} \alpha_i \mathbf{v}_i$, or rather $v_n$ has a representation as a linear combination of the other vectors

#

thank you latex

limber sierra
#

_ outside of math mode fucks things up.

dreamy iron
#

add the grave symbols.

#

it'll clean it up

#

`

limber sierra
#

latex tries to autocorrect since it thinks you forgot to add a $ before the v_n

brisk fractal
#

anyway, it follows directly from the definition, you don't need to prove it really

limber sierra
#

but it doesnt autocorrect it afterwards

dreamy iron
#

anyway, it follows directly from the definition, you don't need to prove it really
@brisk fractal

you must have very generous TAs

desert vessel
#

texit is kind of messed in the sense that dollar signs are bad latex anyway, I much prefer using \( and \[

limber sierra
#

what

#

$ for in-line mathmode is totally fine

#

$$ are more dubious because of spacing reasons

desert vessel
#

yeah, but $$ is not

#

yeah exactly

limber sierra
#

but texit supports \[ too?

#

example: [7x+5]

desert vessel
#

it does? I swear it doesn't

stoic pythonBOT
dreamy iron
desert vessel
#

wow I'm dumb then

limber sierra
#

anyway

#

back on-topic

dreamy iron
#

surround your entire text with it, it'll force the tex to work

limber sierra
#

no

#

that wouldnt have fixed it here

#

the problem was the v_n outside of math mode

#

thats it

desert vessel
#

tex generally works without code mode in markdown

limber sierra
#

latex doesnt like underscores outsidde of math mode

brisk fractal
#

nope didn't fix it

#

whatever

limber sierra
#

sigh

#

people ignore me

#

anyway

desert vessel
#

put $ around $v_n$

limber sierra
#

back on topic

brisk fractal
#

oh shit I see what you mean nami

dreamy iron
#
the definition of linear dependence of a set of vectors is if $0 = \sum \alpha_n \mathbf{v}_n$ for nontrivial $\alpha_n$, i.e. $\mathbf{v}_n = \frac{1}{\alpha_n} \sum_{n \neq i} \alpha_i \mathbf{v}_i$, or rather v_n has a representation as a linear combination of the other vectors
stoic pythonBOT
limber sierra
#

as i said, that does not fix it

#

jesus christr

#

can you read

stoic pythonBOT
limber sierra
#

the problem is the v_n

desert vessel
#

kek

limber sierra
#

latex thinks _

brisk fractal
#

there we go, thank you nami

limber sierra
#

goes in math mode

#

so it tries and add a $

#

because it thinks you forgot it

dreamy iron
#

can you read
@limber sierra i didn't believe you. now i do.

limber sierra
#

aaaanyway

#

this is not linear algebra @wintry steppe

#

and youre interrupting an ongoing discussion anyway

#

okay

#

let me try and

#

get back on topic

#

the nicest proof of the fact you're trying to show ninny is

#

say you have a pair of vectors u, v that is not linearly independent

#

this means that au + bv = 0 for some nonzero a, b

dreamy iron
#

yes.

limber sierra
#

now say we add a new vector to this set, say w

dreamy iron
#

but i have a triplet of vectors

#

okay.

limber sierra
#

then au + bv + 0w = 0

#

since 0w = 0

#

qed

#

oh wait i misread slightly

#

whoops

dreamy iron
#

but i said nothing about a and b's being linearly dependent or not.

limber sierra
#

yeah good point

#

my bad

#

okay so

#

say you have au + bv = w

#

so w is a lin comb of u and v

#

then au + bv - w = 0

#

qed

desert vessel
#

yeah that's how I would prove it

dreamy iron
#

v_1 + v_2 + (av_1 + b_v2)

brisk fractal
#

but yeah, ninny, you're overcomplicating it a bit

desert vessel
#

let $w = av_1 + bv_2$ and there

limber sierra
#

you can give v_1, v_2 any coefficients youd like

dreamy iron
#

okay, i believe i'm adding extra complication.

#

lol

desert vessel
#

you're done

limber sierra
#

because of how linear independence works

#

youre allowed to pick any coefficients

dreamy iron
#

does it work though, as it is?

limber sierra
#

for v_1, v_2

stoic pythonBOT
limber sierra
#

so if you have a linear combination, say av_1 + bv_2

#

then you can just take

#

the coefficients -a and -b

#

-av_1 + -bv_2 + (av_1 + bv_2) = 0

#

by basic algebra

#

if you prefer to write it like that

#

anyway ill skim your proof one sec

#

indeed, you just assert x is not equal to 0 without ever showing it

limber sierra
#

so?

#

i'm assuming that's supposed to be an alpha, not an a

#

but even then

#

z could be 0

#

and then x = -zalpha = 0

dreamy iron
#

yeah, alpha

limber sierra
#

indeed, no matter what you do, x = y = z = 0 will be a solution

#

youre not supposed to reason on that case, youre supposed to show there exist coefficients x, y, z not all 0 that solve 0 = xv_1 + yv_2 + z(av_1 + bv_2)

#

[here a means alpha, b means beta]

dreamy iron
#

but that's trivial solution
we need to find that a zero vector can be found without the trivial solution to show linear-dept.

limber sierra
#

exactly, and you never do that

#

unless you mean youre setting -z to x/alpha

#

but that doesnt work

dreamy iron
#

my starting assumption was only a \neq 0 and b \neq 0

and from that, i found that z, y, x, can all be non-zero and still yeild the zero vector.

limber sierra
#

how? what values should x, y, z take?

#

you never say that

dreamy iron
#

unless you mean youre setting -z to x/alpha
@limber sierra yes?

limber sierra
#

but that doesnt work

#

example:

dreamy iron
#

how? what values should x, y, z take?
@limber sierra they're just elements of a field, any elements.

limber sierra
#

$x\begin{pmatrix}1\0\end{pmatrix} + y\begin{pmatrix}0\1\end{pmatrix} + \frac{-x}{1}\left(1\begin{pmatrix}1\0\end{pmatrix} + 2\begin{pmatrix}0\2\end{pmatrix}\right)$

#

observe that i've set z = x/alpha

#

forgot a negative, whoops

#

whenever texit wakes up...

stoic pythonBOT
limber sierra
#

i forgot to list x, y

#

ugh

#

wait no this is fine

#

and yeah that provides a counterexample to your proposal of

#

setting z = -x/alpha

#

since clearly that sum is (0, -1)

#

which is not the 0 vector

#

even if we assume everything is nice and nonzero

#

(which is true but you dont justify it)

#

as soon as you manipulate the x/a = y/b

#

it is no longer equal to -z

#

indeed, let's think about what you mathematically did to the equation x/a = y/b

#

to get from x/a = y/b to y/x = b/a

#

you multiply both sides by b/x

#

since (x/a)(b/x) = b/a, and (y/b)(b/x) = y/x

#

so if youre doing that to both sides

#

you also have to do it to -z

#

-zb/x = y/x = b/a

#

which... doesnt really tell you much

#

but the point is: you shouldnt have to derive anything (and indeed the thing you derived is false)

#

all you need to do is give an example of values of x, y, z

#

that make xv_1 + yv_2 + z(av_1 + bv_2) = 0 true

#

(of course, your values will depend on the values of a, b)

#

the example i gave is x = -a, y = -b, z = 1

#

then you have:

#

xv_1 + yv_2 + z(av_1 + bv_2) = -av_1 + -bv_2 + (av_1 + bv_2) = 0

#

which is certainly true

#

and this proves linear dependence

#

since we do have a nonzero coefficient

#

(indeed, z is guaranteed to be nonzero, since it's just 1)

dreamy iron
#

the example i gave is x = -a, y = -b, z = 1
@limber sierra

this makes sense.

limber sierra
#

i followed your proposal of setting z = -x/alpha

#

i considered the set {(1, 0), (0, 1)}

#

and then the linear combination 1(1, 0) + 2(0, 1)

#

and then set z = -x/alpha as you suggested

#

but the resulting sum was not (0, 0) the zero vector

#

it was (0, y-2)

#

so clearly your proposal doesnt work

#

oh whoops i just realized i typod

#

replacing that 2 with a 1

dreamy iron
#

yeah, i figured that two was a typo....lemme do some scratch work.

dreamy iron
#

mood

#

i'm trying to recover. that was a lot to take in. i'm still trying to understand what my unjustified step was.

limber sierra
#

your unjustified step was thinking $-z = \frac{x}{\alpha} = \frac{y}{\beta}$ implies $-z = \frac{x}{y} = \frac{\beta}{\alpha}$

stoic pythonBOT
limber sierra
#

heres a simpler example of where this doesnt work:

#

2-1 = 1 = 1

#

but 1=1 becomes 0=0 when we subtract 1 from both sides

#

that dosnt mean 2-1 = 0 = 0

drifting sandal
#

uh.... can linear programming and research op be included here?

dreamy iron
#

no. i never made the leap that -z = b\a

limber sierra
#

ah i misunderstood what you were doing

#

your work is very weird since its unclear what youre actually trying to accomplish

#

let me reread

dreamy iron
limber sierra
#

okay, so it seems your proposal is to set z equal to -x/a, and also equal to -y/b?

#

so youre picking z "after" you pick x and y?

#

but then what should x and y be set to

#

πŸ‘€

dreamy iron
#

anything.

limber sierra
#

but that can't be the case as

#

if you don't set y to the correct value

#

x/a = y/b wont be true

#

indeed, you need to set y = xb/a

#

since then y would be -2x

#

but in any case, all of this work/derivation is a very... indirect way of showing the result

#

the only thing you need to show is that nonzero x, y, z that satisfy this equation exist

#

you dont need to show how you found them or whatever

#

you just need to:

#

#1 - tell us what they are (and that they're nonzero)

#

#2 - show that they make the equation equal to 0

#

now if you set y to the correct value then your proof kind of suffices as step #2?

#

but one needs to go "backwards" through your proof to connect the dots

#

and its not always necessarily clear that these "backwards" manipulations are justified

#

(although in this case, everything you did is invertible so it DOES work, but you never explicitly state that either)

#

a far easier way to show that your proposal for #1 suffices to show #2

#

is to just plug in the values

#

into the equation xv_1 + yv_2 + z(av_1 + bv_2)

#

and show that they make it equal to 0

#

now, dont get me wrong: you're allowed to go the other way to figure out what those values should be

#

(this is very commonly done in epsilon-delta proofs for instance)

#

but you dont need to show us that

dreamy iron
#

the example i gave is x = -a, y = -b, z = 1
@limber sierra

imma go with this. this is short and sweet

#

thank you for your kind time and attention.

undone pumice
#

Please can someone help with the math, not the code, just the math, Thanks!

wintry steppe
#

Would anyone be interested in tutoring me at an hourly rate?

#

Just failed my midterm

half ice
#

Feel free to ask any questions you may have here. We don't condone money transactions over the server

urban karma
#

umm guys, what channel i will ask about quadratic equation

soft burrow
#

or the question channels since that particular channel is busy atm

urban karma
#

oh

#

okay thx

orchid grove
#

Yo guys

#

Does anyone know how to do it?

#

I have been stuck on it for a while now lol

frosty vapor
#

it says quiz tho

#

we cannot help with examinations

mild igloo
#

if I have a matrix a 2x2 and matrix b 3x2 can I do a*b?

half ice
#

Nah fam

soft burrow
#

$b$ would represent, say, a linear transformation $T_b: \bR^2 \to \bR^3$ while $a$ represents a $T_a: \bR^2 \to \bR^2$. The composition $T_a \circ T_b$ is not well defined. Therefore your matrix product $ab$ can't be well defined either

stoic pythonBOT
half ice
#

I like to go with this picture when thinking of matrix multiplications. The only way the matrix multiplication works is if the dot products uses two similarly sized vectors

limber sierra
#

...somehow i'd never realized that one could visualize it like that

#

huh

half ice
#

I never went back

#

Some of my 4th year engineering exams have matrices aligned like this lol

soft burrow
#

my dept's freshman algebra actually motivates matrices by the dot product in some way; start studying rotations in the plane (i.e. R^2) and you'll naturally be led to the rotation matrix

spice storm
#

@soft burrow french?

little frigate
#

That's not french

spice storm
#

spansih

#

spanish

#

@frosty vapor where does it say quiz? on @orchid grove picture?

frosty vapor
#

see the url

orchid grove
#

This isn't a quiz

#

It's a learning activity

#

Where we have multiple attempts and we can answer as many times as possible

#

Want me to prove it

#

Look

#

If u want proof @frosty vapor

frosty vapor
#

oh

#

then sure

#

it's all good sorry to presume

orchid grove
#

All g man

#

Lol I tried the 0 matrix and it didn't work

#

Vector*

soft burrow
#

What kind of quiz tells you that you got the answer right after you input it
many of them actually. even Khan's do iirc
Does anyone know how to do it?
the orth. complement of "a" is all vectors that are orthogonal to "a"

stoic pythonBOT
orchid grove
#

Damn how did u do that so quick lol

#

Ohh ok I got what u mean

#

So u made the variables y and z =0 so the answer doesn't look gross

#

Thanks guys!

#

Also what do those symbols mean

soft burrow
#

So u made the variables y and z =0 so the answer doesn't look gross
yeah, since here we just need one vector

orchid grove
#

Below the solution the bot posted

soft burrow
#

oh, if I react I can e.g. delete the rendered TeX, or delete my original message where I wrote the code

#

for instance I just removed the original messages to clean things up

orchid grove
#

Ohhh ok that's sick

#

I have another one if u got time?

soft burrow
#

I should be studying rn opencry sorry

#

but post your question anyway and see if someone else answers

orchid grove
#

Lol u a graduate student?

soft burrow
#

last year undergrad, taking some grad stuff

orchid grove
#

Ohh ok cya later then and good luck with ur studies!

#

Wait

#

@soft burrow um, it says that it is wrong?

soft burrow
#

uh I might've made some dumb error somewhere. We had x+17w=0 right? so x=-17w, and if w=1 then x=-17. (-17,0,0,1) or (17,0,0,-1) should work

orchid grove
#

Yay thank you!

#

Another question for anyone free, just posted again so u guys don't have to search for it

old flame
#

Hello guys, I am having trouble connecting the proof. I started on both sides, so if $S=S^{m}$, then I'm done. However Im not sure how to show that, and this approach may not even be correct, any help would be appreciated

stoic pythonBOT
old flame
#

<@&286206848099549185>

native rampart
#

What do you think (STS^-1)^2 is?

#

Spoiler: It's not S^2 T^2 S^-2

#

@old flame

old flame
#

Well if it's not that, then it may perhaps be $(STS^{-1})(STS^{-1})=ST^{2}S^{-1}$

stoic pythonBOT
old flame
#

So is it the case that $(STS^{-1})^{n}$, is just the inside composed to itself n times I guess

stoic pythonBOT
native rampart
#

Well if it's not that, then it may perhaps be $(STS^{-1})(STS^{-1})=ST^{2}S^{-1}$
Yes

stoic pythonBOT
native rampart
#

So you get $(STS^{-1})^n=S(T^n)S^{-1}$

stoic pythonBOT
native rampart
#

(Prove that by induction)

old flame
#

Oh.....

#

So are you saying that I need a prove by induction within the question 14 ?

#

And from that result, I could conclude that^^

native rampart
#

Yes

old flame
#

So I guess that result is not given in anyways

#

Sure, thank you

fringe matrix
#

hey

#

quick question: in Matrix, the dimension of its kernel is also the number of "0" as eigenvalues, right?

#

like the power of (lamda) in M(matrix) in the charcristic polynomial

native rampart
#

Yes(No of linearly independent vectors with eigen value 0)

fringe matrix
#

I mean

#

thouse are different things, no?

#

algebric and geometric they are not necessarily equal

native rampart
#

They are always equal

fringe matrix
#

no

#

if the matrix is Diagonalizable they are equal

#

1 2
0 1
The charcristic polynm is (1-x)^2 so "1" is eigenvalue with the power of 2. but the number of lineary independent eigenvectors with eigenvalue "1" is 1

native rampart
#

Nvm

#

My bad

fringe matrix
#

so which is it? :S

#

the kernel is the algebric or geometric multiplicity?

native rampart
#

Geometric

fringe matrix
#

shit..

#

thanks you

old flame
#

Here is my attempt. By theorem 5.10, let $\lambda$ be the eigenvalue of T. Suppose a is an eigenvalue of $P(T)$, then for any $v \in V$, $P(T)v=av$. Then $$P(T)v=[a_0+a_1T+...+a_mT^{m}]v=a_0v+a_1Tv+...+a_mT^{m}(v)]v$$ $$=a_0v+a_1\lambda v+...+a_m \lambda^{m} v=[a_0+a_1\lambda+...+a_m \lambda^{m}]v=P(\lambda)v$$ Thus, $av=P(\lambda)v$, therefore, $P(\lambda)=a$. I think that reversing the steps are also valid, proving both directions.

stoic pythonBOT
native rampart
#

How do you know eigenvectors of P(T) are also eigenvectors of T?

past void
#

Do I have to go through them one by one to prove they fulfill the requirements to be a subspace or is there a more efficient method?

native rampart
#

The efficient method is to show ca+b will be in the space for all vectors a,b and scalars c

past void
#

Okay and then to find a basis for said subspace, in the case of (a) for example, would vectors (2, -1, 5) and (6, -1, 15) be correct?

#

Since any vector in the subspace can be written as a linear combination of those two

native rampart
#

Yes

past void
#

Okayy thanks ❀️

old flame
#

@native rampart I assume that you mean I need to prove the other direction separately then

native rampart
#

No,This direction

#

If P(T)v=av how do you know Tv=lambda v?

old flame
#

because lambda is an eigenvalue of T as I've stated at the start

native rampart
#

I meant to ask,how do you know v should be an eigenvector of T,if it's an eigenvector of P(T)

wintry steppe
#

yea i think this ones impossible

#

impossible to solve

old flame
#

well, is it something like because P(T) are just composite functions in terms of T, so if it is an eigenvector of P(T), then v must be a eigenvector for each term in P(T), which are T's

native rampart
wintry steppe
#

plambda

old flame
#

thanks

sharp totem
#

For a, it is false right? I know that T is injective if A has a pivot in each column

#

But in the case of a 2 x 3 matrix, we have linear dependence

#

So it is not true?

#

And can someone help me with c also?

#

It should be true by definition but not sure how to exactly prove it

waxen jacinth
#

Anyone that can help me with this? He dropped the bomb when my professor uploaded it, I've no clue how to approach it at all xd

hard coral
#

Ask yourself: What is a linear subspace?

native rampart
#

Exact same question was asked above

waxen jacinth
#
  1. Must include the o-th vector
  2. Closed under addition, e.g the sum of two vectors a and b must be enclosed within this subspace
  3. Closed under multiplication, e.g the product of two vectors must be enclosed within the subspace
#

Must've been my friend, we take the same course xD

#

If the answer is in previous messages I can just revise those

hard coral
#

Are you sure about 3?

waxen jacinth
#

No, Im not sure about any of this :')

#

Should be a scalar, not vectors :v

hard coral
#

Right. So if any of those are not fulfilled, it is not a linear subspace

#

As you have to explain your answers, they probably want you to show that you know that :P

waxen jacinth
#

:v

#

At best, I know surface-level theory around these

#

Idk the interactions that arise from the statements

hard coral
#

Well can you express o in W_a?
What happens if you take 2 vectors of the form given in W_a and add them together? Can you rewrite the result that it looks like a vector as given in W_a?
If you take a vector from W_a and multiply it with k can you rewrite it to fit the from given?

waxen jacinth
#

I suppose?

hard coral
#

Do it on paper or however you have to hand in your assignment. That's literally what they want to know. And if you can do all those things, you then have to compute a basis and figure out the dimension

queen gate
#

Hi, I see on Internet that the GCD of two numbers a and b, can be compute using the Thomae's function, like

GCD(a, b) = a f(b/a)

But what is the Thomae's function for a and b with a, b ∈ N ?

limber sierra
#

thomae's function, for a rational number, computes a lowest-terms fraction p/q for b/a, then returns 1/q

#

in practice, this doesnt really help you compute gcds

#

since you still need to compute the lowest-terms representation of b/a

#

which is... basically the same thing as computing the GCD

queen gate
#

Ok, thank you for your explanation.
I work on an exercise in algorithm, and in this exercise, I want to compute the GCD, but do you think there is a formula to compute the GCD easier than the Euler method ?

native rampart
#

You mean euclid?

queen gate
#

yes, sorry πŸ˜…

limber sierra
#

there's the binary GCD algorithm, which is faster than the euclidean algorithm, but harder to do by hand and to implement in an algorithm

#

if you want to compute GCDs by hand, the euclidean algorithm is about as good as you can do.

queen gate
#

I m not going to compute the GCD by hand but with an algorithm... so you think binary gcd algorithm can be good ?

limber sierra
#

what are you hoping to gain out of this algorithm?

#

again, the binary GCD is faster

#

though the reduction won't be DRASTIC

#

up to 50% or so

#

so if the euclidean algorithm was gonna take 10 minutes, the binary algorithm wouldn't suddenly speed that up to something managable

native rampart
#

How is that not drastic?

limber sierra
#

we're talking computational complexity here

#

where efficiency gains can often let you compute orders of magnitudes larger inputs in the same time

queen gate
#

I 'want' an algorithm with the minimum time of computation (and if it's possible, with a minimum complexity)

#

I have only 3.45 sec to compute 784 GCD in the worst case

limber sierra
#

how large are your inputs?

queen gate
#

and a and b, in the worst case are 28

#

a and b between 1 and 28

limber sierra
#

well

#

the fastest algorithm for that would be a lookup table

#

πŸ˜›

queen gate
#

ok

#

hum

#

little error, the worst case is : 69533550916004

limber sierra
#

...okay, then a lookup table is probably impractical

queen gate
#

i think

limber sierra
#

in that case yeah, as far as i know you cant do better than the binary GCD algorithm

queen gate
#

ok

limber sierra
#

i may be incorrect though, not a computer scientist

queen gate
#

I m going to check that

#

okay

limber sierra
#

i know for really large inputs theres a more ring-theoretic algorithm based off lehmer's

#

but that's only faster after like, dozens of digits

queen gate
#

ok

limber sierra
#

and its slower before then

queen gate
#

ok, so I think the binary solution can be a good way

limber sierra
queen gate
#

thx

sharp totem
#

Can someone please help with b?

thorn lichen
#

??

native rampart
#

No c_i is zero

#

@thorn lichen

#

Can someone please help with b?
Just use condition of linear dependence

thorn lichen
#

not sure abou tthis

austere cedar
#

If A is the matrix which rotates all points by theta radians, then the inverse matrix undoes the transformation which a rotation in the opposite direction by theta degrees.

thorn lichen
#

ok thanks

half forge
#

how do you calculate something that is inveetible?

#

invertible

brisk fractal
#

bababooey

native rampart
#

Or just row reduce till you get a row reduced echelon matrix

brisk fractal
#

yeah drake that's the chump easy way out

#

imagine using row reduction when you could calculate a fuckton of subdeterminants

half forge
#

i am reviewing my teachers solution but i dont get how she got it

brisk fractal
#

put the matrix into an augmented matrix with the right side being the identity matrix, then row reduce the left side into reduced echelon form

#

that should transform the right side into the inverse

half forge
#

i am confused at this question

#

how did she get that answer

brisk fractal
#

a matrix has nonzero determinant iff the matrix is invertible

half forge
#

which part a is she talking about?

brisk fractal
#

above, when she calculates the matrix for T in the basis of R^4

half forge
#

you mean this one?

brisk fractal
#

yes

crude oasis
#

does anyone know of a good resource to use for studying bases.... im in my first upper division linear class and was looking for something other than my textbook

half forge
#

i am confused@brisk fractal

#

can you please walk me through the steps

brisk fractal
#

she took the matrix, and calculated the determinant

#

are you confused on how to calculate a determinant

half forge
#

yes, but how did she get these numbers?

#

i put it into the calculator and they got this

brisk fractal
#

yeah I'm unsure what she did to get that determinant

#

I got -4 using row reduction

#

regardless it's still nonzero

half forge
#

oh okay i seee

wide grail
#

Can I get help with 5?

#

This is a sample exam btw thats why it has point numbers

#

What are the dimensions of the vector in R^2 thats being transformed?

stoic pythonBOT
wide grail
#

so this is gonna equal some 3x1 matrix

#

and thats our T?

#

yep got that

#

I just don't understand what any of this means lol

stoic pythonBOT
wide grail
#

oooh

#

so a=2 b=1, c=1 d=-1 e=1 and f=0

#

okay still kinda confused though

#

if the x_1 x_2 vector is in R^2

#

then what is the T matrix in?

stoic pythonBOT
wide grail
#

what the.. never seen that before

#

what does that mean?