1
00:00:00,000 --> 00:00:04,046
[MUSIC]

2
00:00:04,046 --> 00:00:08,720
So now we have some training data and
some desk data.

3
00:00:08,720 --> 00:00:14,105
And what we can do next is #Build,

4
00:00:14,105 --> 00:00:18,935
there's a sub-header here,

5
00:00:18,935 --> 00:00:23,963
#Build the regression model.

6
00:00:25,510 --> 00:00:31,900
So GraphLab Create has many
algorithms pre-implemented for

7
00:00:31,900 --> 00:00:35,000
various machinery and tasks or
you can implement your own if you like.

8
00:00:35,000 --> 00:00:37,616
And during the sequence
of the specialization,

9
00:00:37,616 --> 00:00:41,442
you're gonna write many algorithms
of your own, but for this course,

10
00:00:41,442 --> 00:00:44,379
we're just gonna use
pre-implemented algorithms so

11
00:00:44,379 --> 00:00:47,550
can really try to understand
how machine learning works.

12
00:00:47,550 --> 00:00:49,930
So what you can do with machine learning?

13
00:00:49,930 --> 00:00:56,890
So, we're gonna do is a build
a square foot prediction model.

14
00:00:56,890 --> 00:00:58,480
Just call it sqft_model.

15
00:00:59,590 --> 00:01:06,918
And in GraphLab,
there something called linear regression.

16
00:01:06,918 --> 00:01:13,390
So, linear regression is what we learned,
what we explored and

17
00:01:13,390 --> 00:01:19,700
talked about, and we understood quite well
during the module that Ann was teaching.

18
00:01:19,700 --> 00:01:23,752
And so
if you just say linear _regression.create,

19
00:01:23,752 --> 00:01:28,080
it's going to create a linear
regression model for you.

20
00:01:28,080 --> 00:01:32,174
That creates the traits with
projects GraphLab Create, and

21
00:01:32,174 --> 00:01:37,730
create is also the verb to use to create
various things with GraphLab Create.

22
00:01:37,730 --> 00:01:41,256
So we're now going to create
a linear regression model, and

23
00:01:41,256 --> 00:01:44,450
what I have to give as input
I give it the training data.

24
00:01:45,900 --> 00:01:47,290
So here's the training data.

25
00:01:51,189 --> 00:01:54,780
You have to say what variable
you're trying to predict.

26
00:01:54,780 --> 00:01:56,150
So that's the target.

27
00:01:56,150 --> 00:02:01,790
So I'm gonna say the target
is going to be price.

28
00:02:01,790 --> 00:02:03,420
So I'm trying to predict price.

29
00:02:05,060 --> 00:02:10,760
And the question is what are the features,
what is the input.

30
00:02:10,760 --> 00:02:15,690
So, I'm going to say that the features,
and by the way, you can use top complete

31
00:02:15,690 --> 00:02:18,740
even if [INAUDIBLE] parameters of
function, which is pretty cool.

32
00:02:18,740 --> 00:02:21,020
So, the features is
a list of the features.

33
00:02:21,020 --> 00:02:24,790
If you don't give anything, it will use
all features, all columns of the data.

34
00:02:24,790 --> 00:02:31,520
But here, the only feature input we want
to use is square feet of living space.

35
00:02:32,950 --> 00:02:36,940
And that is a string
that I need to put in, so

36
00:02:36,940 --> 00:02:40,940
I forgot to put it in quotes, so
let me fix that real quick here.

37
00:02:41,980 --> 00:02:46,761
So open quotes,

38
00:02:46,761 --> 00:02:51,203
close quotes.

39
00:02:51,203 --> 00:02:53,070
All right, very good.

40
00:02:53,070 --> 00:02:55,840
So let me just reread that for us.

41
00:02:57,010 --> 00:03:01,959
So sqft_model
graphlab.linear_regression.create on

42
00:03:01,959 --> 00:03:08,270
the training data to predict the target
price using features sqft of living.

43
00:03:08,270 --> 00:03:09,839
Let's go an execute this.

44
00:03:12,830 --> 00:03:17,027
On training data training,
I wrote data twice here.

45
00:03:17,027 --> 00:03:20,910
You see the arrow came out here,
training data and the score data.

46
00:03:20,910 --> 00:03:22,400
So now let's execute.

47
00:03:22,400 --> 00:03:23,710
So what's happening here?

48
00:03:23,710 --> 00:03:31,310
It's done actually, but what we did
was create a linear regression model,

49
00:03:32,570 --> 00:03:37,270
it used a particular algorithm
called Newton's Method.

50
00:03:37,270 --> 00:03:41,510
But GraphLab Create chooses
an algorithm automatically for

51
00:03:41,510 --> 00:03:45,010
you unless you tell it which one to use.

52
00:03:45,010 --> 00:03:50,422
And this algorithm allows you to
make some predictions on the data.

53
00:03:50,422 --> 00:03:54,709
[MUSIC]