#Help with Linear Algebra Problem
284 messages · Page 1 of 1 (latest)
To find if a vector is in the span of another set of vectors, you use an augmented matrix
if the linear system is consistent after doing row reductions, then the vector is in the span of the vectors in your set
Part a looks correct. If you put all of your vectors in a matrix then reduce it to RREF, the basis of the column space will form the basis of the span of your vectors
As a linear algebra teacher, I would prefer you show your row operations.... hopefully your professor doesn't care
would this mean setting A = to
[a]
[b]
[c]
[d]
is there not another way to determine if the vector belongs to the set with the information provided?
so part a actually helps us with part b
by finding a basis, we can set up a simpler linear system, since we now have less vectors describing the same span
so really we only need [a,b | v] where v is the vector we want to check is in the span
actually, your work for a is incorrect
what you found was a basis for the null space
so how would I find a basis for the span
would it be putting it equal to a, b, c, and d?
your pivot columns in your RREF matrix correspond to your basis of the columns space in your original matrix
so in the RREF form, your first and second column are pivot columns
so your basis for the span is the first and second columns in your original matrix (before you reduced it)
right!
so then, instead of having a 5 column augmented matrix in part b, you only need to look at
so then what would be the justification behind this?
$\begin{bmatrix} 1 & 2&1\ 2&5&1\2&2&0\0&1&1\\end{bmatrix}
Cheunk
$\begin{bmatrix} 1 & 2&1\\ 2&5&1\\2&2&0\\0&1&1\\\end{bmatrix}
```Compilation error:```! Missing $ inserted.
<inserted text>
$
l.58 \end{document}
I've inserted something that you may have forgotten.
(See the <inserted text> above.)
With luck, this will get me unwedged. But if you
really didn't forget anything, try typing `2' now; then
my insertion and my current dilemma will both disappear.```
the above should be an augmented matrix
the pivot columns by definition are linearly independent, and row reductions preserves the independence relationship of the columns
so would x3 and x4 be unconstrained variables here?
okay so i don’t think you’re understanding what i mean
look at your RREF matrix
columns 3 and 4 can be written as linear combinations of columns 1 and 2
so we can remove them, and the resultant set will still have the same span
thus we have a linearly independent spanning set aka a basis
ohhhhh i get it now ty
so by linear combination that means that the columns can be multiplied by x3 and x4 right?
a linear combination is a sum of any scalar multiple of the vectors in the set
so in your RREF matrix column 3 is a linear combination of columns 1 and 2 because if we multiply column 1 by -1 and add it to column 2, we get column 3
So do you have to do those calculations to determine whether they are linear combinations or can you figure it out looking at the equations i wrote down below the matrices?
the equations are irrelevant
you only need to look at the RREF
i would suggest you look over these definitions in your textbook
does that mean the only way to determine if there are linear combinations is to look at the RREF and check each vector?
Kind of
by reducing the matrix, we can find all of the pivot columns
those are the columns that have a ‘1’ in one spot, and ‘0’ at every other spot in the column
if we can identify the pivots, we know all of the other columns are in the span of the pivots
so we can remove them without changing the span
does this hold true everytime?
if you have a textbook, you should look at the definitions for pivot, linear independence, span, and linear combination
yes. if you have a set of vectors, put them in a matrix, then reduce it to RREF, the columns in the ORIGINAL matrix, which are in the same spot as the pivot columns in the RREF will always be a linearly independently spanning set
i think i get it now
so basically just reduce to RREF and take the columns from the original matrix from the same positiion as the pivot columns
right! then when you do part b, you use an augmented matrix with those columns
the basis columns or the pivot columns?
if the matrix is consistent, then the answer is yes, else that vector is not in the span
or just A
the basis columns
oh
see here
whre does 1 1 0 1 come from?
that is the one the problem is asking you to see is in the span!
so what does that mean as a linear system?
exactly!
so that vector cannot be in the span of the others
you could also have done that by putting all of the vectors in a matrix and augmenting the final vector
but part a allows us to simplify the row operations by getting rid of 2 of the columns
so putting A equal to 1 1 0 1?
yes that also gave 0=1
but couldnt you also check by putting x1=1, x2=1, x3=0, x4=1 into the equations x1-x3+2x4=0 and x2+x3-x4=0?
sure, but this does not use any linear algebra techniques
so its a valid way to check but it just doesnt use linear algebra
but it gives the same results correct?
yes. your professor would probably take points off if you did it that way though 😂
oh ok lol
would multiplying the matrix A and the given vector and seeing if it equals zero also work?
no. if you multiply Av you will just get another vector. it would not tell you if the vector was in the span of the columns of the matrix
if you multiply Av and get zero it would tell you that v is in the null space of A though
Oh i see
Ok then what would be the justification behind augmenting matrix A and the given vector to see if it is consistent?
If it is consistent after you do the row reductions
then you can solve the linear system to find a solution
this solution will be a set of scalars that you scale your vectors by to make the vector as a linear combination
this example explains it well
So could you say that because the system is inconsistent after doing row operations, there are no scalars that can give you the vector provided
i dont know how to word it
yeah that’s close enough
or you could say
since the matrix is inconsistent, the is no c1,c2 such that
v = c1 y1 + c2 y2
where v is (1 1 0 1)
and y1 and y2 are your vectors from part a
remember, the span is the set of ALL linear combinations
so if there is no such linear combination, the vector cannot be in the span
linear combinations of y1 and y2?
so
if you have a set of vectors
say S = {y1, y2, y3}
then the span of S is all of the linear combinations of y1, y2, y3
for example
-2 y1 + 3y2 -πy3 is in the span
so is
2y1 - y3
it’s ANY sum of ANY scalar multiple of the vectors in the set
so for the span of S it would by any scalar multiples of the 4 vectors in S?
right! that’s the span of S
so what you’ve done is ask
can i find scalars multiples of these vectors which add to (1 1 0 1)
and so if you set S equal to the vector you want if you dont get a solution then there are no scalar multiples that can get you to that vector
exactly!!
and what part A did was get rid of the extra vectors in the set that weren’t “making the span bigger”
the set we have from the problem statement and the set we got as an answer for part a have the same span
ok that makes much more sense now
so the justification would be the same for the basis vectors and all 4 vectors right?
since they are basically the same
right. the benefit of removing the vectors we did in part a though is that we have a lot less row operations to do in part b
there are 2 less columns to worry about
so its basically just to make it easier
yes
its equal to zero?
oh right
so that way you don’t need row reductions, just matrix vector multiply
also, to help you understand span
i want you to pick any x3 and x4 from this part and calculate what vector you get from that
then multiply that vector by A on the left
if you did that problem right, you will ALWAYS get 0
looks good to me!
the benefit of a basis is
any vector in the span of the basis vectors can be written as a unique linear combination of the basis vectors
if you have a set that spans but isn’t linearly independent, you can represent vectors in a lot of different ways
but there’s only one linear combination for each vector if you have a basis
so your intuition was correct
to do part d, you could either use an augmented matrix with your basis vectors as columns, or multiply the vector by the matrix A and see if you get 0
just to make sure, part c is asking for a spanning set for N(A)
spanning set and basis are different right?
ah good point! a basis is a spanning set which is linearly independent
so a basis is a spanning set
but not necessarily the other way around
remember a basis of a space H is a set that:
a. is linearly independent
b. spans H
if you get rid of a. you only have a spanning set
does that mean a spanning set is only linearly dependent?
no it only spans
If it spans the space, it’s a spanning set
if it is linearly independent, it is a linearly independent set
if it is both, it’s a basis
no there can be many!
you can have an infinite number of bases of ℝ² for instance
{(1 0),(0 1)} is the “standard basis”
{(-1 2) , (-4 -7) } is also a basis
(and therefore also both spanning sets)
so would {(-1, 2), (-2, 4)} be a spanning set but not a basis because its linearly dependent
hmm well not quite
it would span something, but not ℝ²
{(-1 2), (-2 4), (-1 -1)} would be a spanning set of ℝ² but not linearly independent
notice that you can only make the span bigger or keep it the same by adding vectors
so a better way to find a spanning set that isn’t linearly independent would be to add a vector that is already in the span of the stuff we have in the basis (like what i did)
ah ok
wait sho how did you get {(-1 2) , (-4 -7) } is also a basis
did you just choose two random linearly independent vectors?
yes! we have a shortcut for sets of 2 vectors
if one is not a scalar multiple of the other, the set is linearly independent
if i multiply the first vector by 4 to make them agree in the first coordinate, that forces my second coordinate to be 8
8 is not equal to -7 so they are linearly independent
so what made {(-1 2) (-2, 4)} not a spanning set?
well (-2,4) is in the span of (-1 2)
so without decreasing the span, we can consider the set containing only (-1 2)
now the vector (1 1) is in ℝ²
but it is not in the span of (-1 2)
what about (-1 2) (-2 4) and (-3 6)?
ah! these are good questions
both (-2 4) and (-3 6) are in the span of (-1 2)
so we can throw then both out without decreasing the span! still not a spanning set of ℝ²
so we have a nice theorem here
if we have a linearly independent set of n many vectors
it will always be a basis of ℝ^n
so if we have 3 linearly independent vectors from ℝ³
they will form a basis of ℝ³
so then what defines a spanning set of ℝ²?
A set which can make any vector in ℝ²
as a linear combination
think of the set of all linear combinations of (1 0) and (0 1)
they look like
a(1 0) + b(0 1)
if we put those together
a(1 0) + b(0 1) = (a b)
well since i can choose any a and any b, i can get anything in ℝ²!
now lets add a new vector (1 1)
i can make that already out of my basis
so now my set is linearly dependent
but now the linear combinations look like
a(1 0) + b(0 1) + c (1 1) = (a + c b + c)
but if i just make c = 0, i have the same thing as before!
so this set is still a spanning set, but no linearly independent (so not a basis)
ok i think i kinda get it but im still confused as to why {(-1, 2), (-2, 4)} isnt a spanning set
is it because one of the vectors spans the other
and by spanning the other vector does that mean they are scalar multiples?
so it is a spanning set for the span of either of the individual vectors
but it is not a spanning set of ℝ²
because there are vectors that we can’t make out of those 2 vectors
everything in the span of those vectors looks like (-a 2a)
and there are definitely things that don’t look like that
what would be an example of a vector that we cant make?
(1 1)
oh i see now
you can check using an augmented matrix. it will be inconsistent
so then how do you know if two vectors will be a spanning set of ℝ²
if they arent scalar multiples are they always a spanning set of ℝ²?
yeah that theorem i told you earlier applies
any 2 linearly independent vectors will be a basis
and in ℝ³ any linearly independent set of 3 vectors will form a basis, etc etc
the fastest way to check if a set of vectors is a basis is to first count that you have enough
then put them all in a matrix
if the matrix reduces to the identity, they form a basis
ah i see now
and just to be sure you can have more than 2 vectors that are a spanning set of ℝ² correct?
yes!
this example is a spanning set that
has more than 2
ok perfect
also back to part d, you said that i can check if the vector belongs to the null space by augmenting it with the basis vectors
i did that but got an inconsistent system even though i multiplied A and x and got zero
did you do this with the basis vectors of the null space?
wouldnt the basis vectors be the same for part a and c?
no! the basis from c gives you the basis for all of the vectors that get mapped to 0 by A
the basis from a gives you a basis for all of the vectors that are linear combinations of the columns of A
so basically, the basis from a gives you a basis to find all b such that Ax = b
the basis from c gives you a basis to find all x such that Ax = 0
so in this case is the spanning set the same as the basis for part c?
i was just confused because using the equations for the null space gives you the basis vectors but for part a the equations gave the x3 and x4 which were removed to get the basis vectors
yes they are different processes
it comes from the fact that the null space is a subspace of the domain but the column space is a subspace of the codomain
Ok and very last question, can I check if the vector belongs to the null space by checking if it matches with the equations I derived like I could do in part a and b?
this setup
what about for the equations x1-x3+2x4=0 and x2+x3-x4=0?
i just want to know if it works or not
you should avoid doing that in general
but you could do it with your basis vectors from the null space instead of your vectors from part a
so as you’ve written it that would not check if it’s in the null space
instead it would be
using the numbers from the vectors in the basis you found for part c
I get
1 0 1
0 1 1
0 0 0
0 0 0
so as long as its consistent even if its something like
1239 9412 2
0 241 429849
0 0 0
0 0 0
it would belong to the null space?
yup
ok awesome
you're welcome! you were asking good questions so it was nice to help you
i have no idea
o lol
look under the #❓how-to-get-help channel maybe?
i think i figured it out
before i close it is it cool if i add u if i have anymore questions? its cool if u dont want to tho
cool
sure go ahead. i can't guarantee i can help you quickly lol i was at the gym this whole time and answering between sets lol
o dang lol, mb if i bothered u
thanks for the help anyways have a good rest of your night