WEBVTT
00:00:07.065 --> 00:00:07.920
PROFESSOR: Hi there.
00:00:07.920 --> 00:00:09.540
Welcome back to recitation.
00:00:09.540 --> 00:00:12.870
In lecture, you've been learning
about when vectors are linearly
00:00:12.870 --> 00:00:17.550
independent, when they span
the space, what a basis is,
00:00:17.550 --> 00:00:19.790
what a dimension of
a vector space is.
00:00:19.790 --> 00:00:23.630
And the problem for today
is exactly about that.
00:00:23.630 --> 00:00:26.500
We have a vector
space that is given.
00:00:26.500 --> 00:00:28.780
It's spanned by
these four vectors.
00:00:28.780 --> 00:00:32.009
And we're asked to find the
dimension of that vector space
00:00:32.009 --> 00:00:33.190
and the basis for it.
00:00:35.700 --> 00:00:39.270
Well, why don't you hit pause
on the video, and work on it
00:00:39.270 --> 00:00:39.770
for a while.
00:00:39.770 --> 00:00:42.510
And I'll come back in a little
bit to help you out with it.
00:00:51.120 --> 00:00:51.900
We're back.
00:00:51.900 --> 00:00:54.900
Let's work on it.
00:00:54.900 --> 00:01:00.310
So we need to find the
dimension and the basis.
00:01:00.310 --> 00:01:01.970
Remember what the dimension is?
00:01:01.970 --> 00:01:03.570
It's simply the
number of vectors
00:01:03.570 --> 00:01:05.590
in a basis for the vector space.
00:01:05.590 --> 00:01:07.680
So actually, the
problem is backwards.
00:01:07.680 --> 00:01:09.460
We want to find the
basis for the space
00:01:09.460 --> 00:01:11.840
first, and then
find the dimension.
00:01:11.840 --> 00:01:19.700
I'll write "first"
here and "second" here.
00:01:19.700 --> 00:01:22.270
So we want to find a basis
for the vector space spanned
00:01:22.270 --> 00:01:24.040
by these four vectors.
00:01:24.040 --> 00:01:26.040
So you might be
tempted to just say
00:01:26.040 --> 00:01:27.680
that a basis for
this vector space
00:01:27.680 --> 00:01:31.410
is those four vectors, because
they span the vector space.
00:01:31.410 --> 00:01:34.350
But there's another thing
that a basis has to satisfy.
00:01:34.350 --> 00:01:36.920
And it is the elements
of the basis have
00:01:36.920 --> 00:01:38.760
to be linearly independent.
00:01:38.760 --> 00:01:40.340
We don't know that these are.
00:01:40.340 --> 00:01:42.636
So we have to check.
00:01:42.636 --> 00:01:44.010
How do we check
that four vectors
00:01:44.010 --> 00:01:45.670
are linearly independent?
00:01:45.670 --> 00:01:47.800
Well, there's a couple
of different ways.
00:01:47.800 --> 00:01:50.610
But here's what
we're going to do.
00:01:50.610 --> 00:01:54.340
We're going to put these
vectors as rows of a matrix.
00:01:54.340 --> 00:01:57.070
And then we'll do elimination.
00:01:57.070 --> 00:02:00.710
And when we get to the end,
the rows that have pivots
00:02:00.710 --> 00:02:02.570
are the independent ones.
00:02:02.570 --> 00:02:04.220
So let's do that.
00:02:06.760 --> 00:02:25.740
1, 1, -2, 0, -1; 1, 2, 0, -4, 1;
0, 1, 3, -3, 2; 2, 3, 0, -2, 0.
00:02:30.480 --> 00:02:32.730
By now you must have done
elimination a million times,
00:02:32.730 --> 00:02:36.310
so I'll go a little bit faster.
00:02:36.310 --> 00:02:48.864
1, 1, -2, 0, -1; 0,
1, 0, +2, -4, 2--
00:02:48.864 --> 00:02:56.670
this one's already done-- 1,
3, -3, 2; 2 minus 2, 3 minus 2,
00:02:56.670 --> 00:03:02.170
0 plus 4, -2, and 2.
00:03:05.457 --> 00:03:06.040
One more step.
00:03:08.740 --> 00:03:21.146
1, 1, -2, 0, -1-- all these
are done-- 0, 1, 2, -4, 2; 0,
00:03:21.146 --> 00:03:30.491
1 minus 1 is 0, 3 minus 2 is 1,
-3 plus 4 is 1, 2 minus 2 is 0,
00:03:30.491 --> 00:03:37.402
1 minus 1 is 0,
2, 2 again, and 0.
00:03:37.402 --> 00:03:39.280
Well, you can see
where this is going.
00:03:39.280 --> 00:03:41.760
In the next step, this
row is going to disappear.
00:03:44.748 --> 00:03:59.120
1, 1, -2, 0 -1; 0, 1, 2, -4, 2;
0, 0, 1, 1, 0; 0, 0, 0, 0, 0.
00:03:59.120 --> 00:03:59.954
All right.
00:03:59.954 --> 00:04:01.120
We're done with elimination.
00:04:03.890 --> 00:04:05.300
So let's circle our pivots.
00:04:08.934 --> 00:04:10.470
All right, here are our pivots.
00:04:10.470 --> 00:04:14.040
We have three pivots.
00:04:14.040 --> 00:04:17.340
And so these three rows
are linearly independent.
00:04:19.950 --> 00:04:24.140
And in fact, these rows
still span the same space
00:04:24.140 --> 00:04:25.801
that the initial four rows did.
00:04:25.801 --> 00:04:28.050
Because when you do elimination,
all that you're doing
00:04:28.050 --> 00:04:31.380
is recombining
your rows by doing
00:04:31.380 --> 00:04:32.560
linear combinations of them.
00:04:32.560 --> 00:04:36.310
So, for example, your first
row stayed the same throughout.
00:04:36.310 --> 00:04:41.850
Your second row was replaced
by row 2 minus row 1.
00:04:41.850 --> 00:04:45.310
But it's really still
spanning the same space.
00:04:45.310 --> 00:04:46.330
And it goes on.
00:04:46.330 --> 00:04:48.640
And then the fourth row,
it turns out, was useless.
00:04:48.640 --> 00:04:50.590
You only needed the first three.
00:04:50.590 --> 00:04:57.295
So the elements for a basis--
well, it will be these three.
00:04:57.295 --> 00:04:58.170
So let me write that.
00:05:01.170 --> 00:05:18.640
Basis [1, 1, -2, 0, -1],
2, 2, and [0, 0, 1, 1, 0].
00:05:18.640 --> 00:05:22.440
Could you have used
the first three rows?
00:05:22.440 --> 00:05:24.230
Yes, you could.
00:05:24.230 --> 00:05:25.640
You can't always do that.
00:05:25.640 --> 00:05:28.390
Sometimes in elimination,
you have to switch rows,
00:05:28.390 --> 00:05:30.750
because there's a 0
where a pivot should be.
00:05:30.750 --> 00:05:33.380
When that happens, you
have to use these three,
00:05:33.380 --> 00:05:35.310
or you have to keep
track of which row
00:05:35.310 --> 00:05:39.200
you switched to go back
and use the initial ones.
00:05:39.200 --> 00:05:41.110
But it's really safe
to use these ones.
00:05:41.110 --> 00:05:43.880
And also they're simpler than
the ones that you started with,
00:05:43.880 --> 00:05:45.820
so why not?
00:05:45.820 --> 00:05:47.840
The other question
that we had was:
00:05:47.840 --> 00:05:49.740
what is the dimension
of the vector space?
00:05:49.740 --> 00:05:52.780
Well, this is the easy part.
00:05:52.780 --> 00:05:58.100
The dimension of the
vector space is 1, 2, 3.
00:05:58.100 --> 00:05:59.620
And that solves the problem.
00:05:59.620 --> 00:06:03.530
But there's one more thing
that I want to tell you.
00:06:03.530 --> 00:06:05.210
I told you that there's
a couple of ways
00:06:05.210 --> 00:06:08.950
to find out which of
those four vectors
00:06:08.950 --> 00:06:10.480
are linearly independent.
00:06:10.480 --> 00:06:12.990
And the one that I used
was I put them into rows
00:06:12.990 --> 00:06:16.920
and performed elimination
and picked out the rows
00:06:16.920 --> 00:06:19.050
that have pivots on them.
00:06:19.050 --> 00:06:23.380
Another way to do it is to
write the initial vectors
00:06:23.380 --> 00:06:26.840
as columns of the matrix and
then perform elimination.
00:06:26.840 --> 00:06:28.880
That also works
and, as you know,
00:06:28.880 --> 00:06:31.870
because you're only working with
the transpose of this matrix,
00:06:31.870 --> 00:06:34.990
you get the same
number of pivots.
00:06:34.990 --> 00:06:40.250
Let's go over here where I have
the same-- well, the transpose
00:06:40.250 --> 00:06:42.260
of that matrix-- I
have the same vectors,
00:06:42.260 --> 00:06:44.100
but now written as columns.
00:06:44.100 --> 00:06:46.740
My four initial vectors
written as columns.
00:06:46.740 --> 00:06:49.440
And here I have
performed elimination.
00:06:49.440 --> 00:06:55.466
And this is the final result.
Let me circle the pivots again.
00:06:55.466 --> 00:06:57.870
Here they are,
three, which is going
00:06:57.870 --> 00:07:01.210
to give me three linearly
independent vectors
00:07:01.210 --> 00:07:04.060
and hence, three elements
of the basis, and hence,
00:07:04.060 --> 00:07:06.350
dimension 3 for
the vector space.
00:07:06.350 --> 00:07:11.040
But I could no longer
use these three columns
00:07:11.040 --> 00:07:15.040
as elements of the basis,
because when I did elimination,
00:07:15.040 --> 00:07:17.500
I changed the column space.
00:07:17.500 --> 00:07:20.410
So the column space
of this matrix
00:07:20.410 --> 00:07:22.650
is not the same as
the column space
00:07:22.650 --> 00:07:24.290
of the eliminated version.
00:07:24.290 --> 00:07:27.410
So I cannot use these.
00:07:27.410 --> 00:07:29.550
In fact, if you
look at them, you
00:07:29.550 --> 00:07:31.550
can probably understand
that they're not
00:07:31.550 --> 00:07:34.790
going to span the
same space as these.
00:07:34.790 --> 00:07:37.040
Because all that I
have down here are 0's,
00:07:37.040 --> 00:07:39.960
and I get a lot more than
just 0's in the last two
00:07:39.960 --> 00:07:41.840
entries of the vectors.
00:07:41.840 --> 00:07:45.260
So what I need to
do is-- the pivots
00:07:45.260 --> 00:07:48.310
are in the first, second,
and third columns.
00:07:48.310 --> 00:07:53.550
I need to use these three
columns as my basis.
00:07:53.550 --> 00:07:57.420
I'll just write basis down here.
00:07:57.420 --> 00:07:58.790
And that will work too.
00:07:58.790 --> 00:08:01.200
So see, I have produced
two different bases
00:08:01.200 --> 00:08:04.190
for the same vector space,
which is totally fine.
00:08:04.190 --> 00:08:06.420
You can pick the
basis that you prefer.
00:08:06.420 --> 00:08:07.100
All right.
00:08:07.100 --> 00:08:08.040
We're done.
00:08:08.040 --> 00:08:09.398
Thank you.