1
00:00:00,000 --> 00:00:03,499
[MUSIC]

2
00:00:03,499 --> 00:00:08,013
Okay, so far we've assumed that the only
feature relevant to the value of my house,

3
00:00:08,013 --> 00:00:10,620
is the square feet of the house.

4
00:00:10,620 --> 00:00:13,400
But I dig into the data
set a little bit more, and

5
00:00:13,400 --> 00:00:17,670
I look at a house that's supposed
to be very similar to mine.

6
00:00:17,670 --> 00:00:20,720
It's a house that has very
similar square footage.

7
00:00:20,720 --> 00:00:25,890
And so, this house is definitely making an
influence on what the predictions are for

8
00:00:25,890 --> 00:00:27,010
my house.

9
00:00:27,010 --> 00:00:31,050
But I look at it and I say, and
I look at the specific listing, and

10
00:00:31,050 --> 00:00:33,650
it shows that that house
only had one bathroom.

11
00:00:33,650 --> 00:00:36,650
It's actually quite a big house,
only one bathroom.

12
00:00:36,650 --> 00:00:39,990
And I say that's really,
really not comparable to my house,

13
00:00:39,990 --> 00:00:41,699
which has three bathrooms.

14
00:00:43,160 --> 00:00:46,810
So that the value of that house,
that house sales price

15
00:00:46,810 --> 00:00:50,210
really shouldn't be indicative of
what my house sales price should be.

16
00:00:52,140 --> 00:00:55,330
So instead what we can think about
doing is adding more features.

17
00:00:55,330 --> 00:00:59,050
So instead of just looking at
the relationship between square feet and

18
00:00:59,050 --> 00:01:03,120
price, we can add number of bathrooms.

19
00:01:03,120 --> 00:01:03,630
And now for

20
00:01:03,630 --> 00:01:07,100
each one of the listings that I looked at
before, I'm gonna have to go through and

21
00:01:07,100 --> 00:01:11,770
record how many square feet that house
had, and the number of bathrooms.

22
00:01:11,770 --> 00:01:15,470
And I'm gonna plot each of
these points in the 3D space.

23
00:01:15,470 --> 00:01:18,390
Okay?
So it's this hyper cube of square

24
00:01:18,390 --> 00:01:21,710
feet versus bathrooms versus price.

25
00:01:21,710 --> 00:01:26,130
And now instead of fitting a line to the
data, if I'm thinking about just a very

26
00:01:26,130 --> 00:01:30,360
simple model,
I can think about fitting a hyper plane.

27
00:01:30,360 --> 00:01:33,370
Okay, so
it's just a slice through the space.

28
00:01:33,370 --> 00:01:34,010
We're here.

29
00:01:34,010 --> 00:01:38,690
This is the equation of the hyper plane,
and this is the equation of this plane.

30
00:01:38,690 --> 00:01:43,120
So we have w0, which is our intercept,
just where this plane lives up and

31
00:01:43,120 --> 00:01:45,110
down on the y-axis.

32
00:01:45,110 --> 00:01:47,280
And we have w1 times
the number of square feet, and

33
00:01:47,280 --> 00:01:50,260
w2 times the number of bathrooms.

34
00:01:50,260 --> 00:01:51,730
But a question is where do we stop?

35
00:01:51,730 --> 00:01:56,340
Do we just want to include the number
of bathrooms as our additional feature?

36
00:01:56,340 --> 00:01:58,546
There are lots of things we
could think about including.

37
00:01:58,546 --> 00:02:02,244
We could think about in addition to
our square feet, number of bathrooms,

38
00:02:02,244 --> 00:02:05,882
there's the number of bedrooms,
the lot size, how old the house is, and

39
00:02:05,882 --> 00:02:07,074
the list goes on and on.

40
00:02:07,074 --> 00:02:10,893
In terms of different properties of
the house that could be influential in

41
00:02:10,893 --> 00:02:12,193
assessing it's value.

42
00:02:12,193 --> 00:02:17,660
But we're gonna actually hold off on this
question of looking at which features

43
00:02:17,660 --> 00:02:23,220
are important for this regression task,
until we get to the regression course.

44
00:02:23,220 --> 00:02:25,820
So go to the regression course
to learn more about this topic.

45
00:02:25,820 --> 00:02:29,979
[MUSIC]