1
00:00:00,133 --> 00:00:00,500
All right.

2
00:00:00,500 --> 00:00:03,033
So let's do this. Let's create
a new code cell.

3
00:00:03,033 --> 00:00:07,133
And now the first thing that we have to do
is called the matplotlib

4
00:00:07,133 --> 00:00:10,133
dot pyplot module,
which has a shortcut plt.

5
00:00:10,800 --> 00:00:15,266
Then from this module we're going to call
first the scatter function

6
00:00:15,700 --> 00:00:19,800
which will allow to first,
you know, display on a 2D plot

7
00:00:20,066 --> 00:00:24,166
the different points of coordinates
containing the real results.

8
00:00:24,166 --> 00:00:29,466
You know, the real position levels
going from 1 to 10 and the real salaries.

9
00:00:29,466 --> 00:00:30,933
So that's what we are plotting first.

10
00:00:30,933 --> 00:00:33,066
And then we'll plot the predictions.

11
00:00:33,066 --> 00:00:36,400
And so well here
we have to input for the coordinates

12
00:00:36,400 --> 00:00:40,300
of these real points containing
the real position levels in real salaries.

13
00:00:40,600 --> 00:00:43,233
And on the x axis we have
of course the position levels.

14
00:00:43,233 --> 00:00:45,833
Therefore the x coordinates will be x.

15
00:00:45,833 --> 00:00:48,533
You know
all the position levels contained in x.

16
00:00:48,533 --> 00:00:51,666
And then on the y
axis will have the salaries the real ones.

17
00:00:51,900 --> 00:00:54,300
And therefore the y coordinates will be.

18
00:00:54,300 --> 00:00:56,266
Y'all right.

19
00:00:56,266 --> 00:00:58,766
And then remember we need to add a color

20
00:00:58,766 --> 00:01:01,766
and we'll choose like last time red.

21
00:01:01,800 --> 00:01:03,833
Perfect. So that's for the real results.

22
00:01:03,833 --> 00:01:05,900
And now we're going
to plot the predictions.

23
00:01:05,900 --> 00:01:09,466
And to do that
well we're going to call PLT again

24
00:01:09,466 --> 00:01:14,200
matplotlib.pyplot from which
we're going to call the plot method.

25
00:01:14,200 --> 00:01:17,833
Because this time we're actually going
to plot that regression

26
00:01:17,833 --> 00:01:20,833
line, you know, in
blue like in simple linear regression.

27
00:01:20,866 --> 00:01:22,400
And then for polynomial regression,

28
00:01:22,400 --> 00:01:25,866
you will see that it's not a line,
but it will be actually a curve.

29
00:01:25,866 --> 00:01:28,700
And still
we will use the plot function for that.

30
00:01:28,700 --> 00:01:33,900
So again we have to enter the coordinates
of the different points of this line.

31
00:01:34,133 --> 00:01:35,533
And for these coordinates

32
00:01:35,533 --> 00:01:39,600
well first the x coordinates are still x
the position levels.

33
00:01:39,866 --> 00:01:43,666
But then the y coordinates will be
of course the predicted salaries.

34
00:01:43,700 --> 00:01:45,900
You know, instead of the real salaries.

35
00:01:45,900 --> 00:01:48,933
And to get them
well we simply need to call our line

36
00:01:49,366 --> 00:01:52,000
req object not line rec two.

37
00:01:52,000 --> 00:01:54,700
You know, line rec two
is for the polynomial regression model.

38
00:01:54,700 --> 00:01:57,200
Line is for the linear regression model.

39
00:01:57,200 --> 00:02:00,566
And so Lin rec from which
we're going to call that predict

40
00:02:00,866 --> 00:02:04,566
method applied of course to X right.

41
00:02:04,733 --> 00:02:07,900
Containing the position levels
of the matrix of features x.

42
00:02:08,433 --> 00:02:10,633
All right. And then we add a color.

43
00:02:10,633 --> 00:02:13,633
And like last time
we'll choose the color blue.

44
00:02:14,033 --> 00:02:15,333
All right blue.

45
00:02:15,333 --> 00:02:18,333
And then we'll just improve or enhance

46
00:02:18,333 --> 00:02:21,566
our graph by adding a title
an x label and a y label.

47
00:02:21,566 --> 00:02:24,733
And finally displayed
so you know how to do this

48
00:02:24,733 --> 00:02:28,600
we call first plt then the title function.

49
00:02:28,966 --> 00:02:32,366
So here we're going to add a funny title
like and quote.

50
00:02:32,400 --> 00:02:37,266
Of course truth or fluff right.

51
00:02:37,266 --> 00:02:40,266
This is the simple scenario
I invented for this case study.

52
00:02:40,633 --> 00:02:44,433
And then we're going to precise that
it's linear regression.

53
00:02:44,433 --> 00:02:48,633
You know the linear regression model
okay good good title.

54
00:02:48,833 --> 00:02:50,466
Now an X label.

55
00:02:50,466 --> 00:02:53,533
So here again plt dot x label.

56
00:02:53,966 --> 00:02:55,166
And we'll choose.

57
00:02:55,166 --> 00:03:00,633
You know something
simple like position level okay perfect.

58
00:03:00,633 --> 00:03:02,600
Then a y label.

59
00:03:02,600 --> 00:03:05,400
So I'm calling the y label function.

60
00:03:05,400 --> 00:03:08,400
And we're going to enter salary
because on the

61
00:03:08,400 --> 00:03:12,066
y axis are all the salaries
the real ones or the predicted ones.

62
00:03:12,633 --> 00:03:19,233
And finally remember that we have to show
the graph by using this show function.

63
00:03:19,566 --> 00:03:20,733
Perfect.

64
00:03:20,733 --> 00:03:21,466
And there we go.

65
00:03:21,466 --> 00:03:24,400
We're already ready to visualize
the linear regression results.

66
00:03:24,400 --> 00:03:25,133
Are you ready?

67
00:03:25,133 --> 00:03:28,800
I'm going to close this
so that we can see it very well.

68
00:03:29,300 --> 00:03:31,566
All right
let's make sure I didn't make any mistake.

69
00:03:31,566 --> 00:03:33,266
Yeah it seems all good.

70
00:03:33,266 --> 00:03:36,266
So now let's press play and here we go.

71
00:03:36,366 --> 00:03:39,300
This is the linear regression results.

72
00:03:39,300 --> 00:03:41,400
Let's scroll down a bit.

73
00:03:41,400 --> 00:03:42,033
All right.

74
00:03:42,033 --> 00:03:47,533
So first of all to recap
the red points here are the real salaries.

75
00:03:47,533 --> 00:03:52,500
You know going from zero to
I think 1 million for the CEO.

76
00:03:52,500 --> 00:03:53,600
Yes okay.

77
00:03:53,600 --> 00:03:57,133
So these are all the real salaries
of this column okay.

78
00:03:57,333 --> 00:03:59,766
And then the blue line is of course

79
00:03:59,766 --> 00:04:02,900
the regression line
containing all our predictions.

80
00:04:02,900 --> 00:04:05,900
But the predictions of the linear
regression model.

81
00:04:06,266 --> 00:04:09,600
So first of all we can see that
indeed the linear

82
00:04:09,600 --> 00:04:13,933
regression model
is not well adapted to this data set.

83
00:04:13,933 --> 00:04:17,500
Because indeed, remember that 
in simple linear regression,

84
00:04:17,500 --> 00:04:20,900
it was well adapted because each time,
you know, for each of the values

85
00:04:20,900 --> 00:04:25,100
of the features, well, the prediction
was close to the real result.

86
00:04:25,500 --> 00:04:29,366
But here for many position levels,
you know, for many values

87
00:04:29,366 --> 00:04:33,466
of the feature, well, the prediction is
far from the real result here.

88
00:04:33,466 --> 00:04:36,166
For example,
it is quite far from the real salary.

89
00:04:36,166 --> 00:04:40,666
Imagine if we wanted to use that model
to predict if someone is saying

90
00:04:40,666 --> 00:04:43,733
the truth, or bluffing about a salary
that is contained here.

91
00:04:43,866 --> 00:04:44,700
Well, you know,

92
00:04:44,700 --> 00:04:48,500
we would have offered a way higher salary
then we would be supposed to.

93
00:04:48,500 --> 00:04:50,833
So that would not be the best negotiation.

94
00:04:50,833 --> 00:04:51,200
Okay.

95
00:04:51,200 --> 00:04:54,866
So that linear regression
model is not well adapted and same here.

96
00:04:54,866 --> 00:04:56,400
It is far from the prediction.

97
00:04:56,400 --> 00:04:59,100
Same here, far from the prediction
here. It's okay.

98
00:04:59,100 --> 00:05:02,433
But you know that's only for two points
two salaries.

99
00:05:02,733 --> 00:05:04,933
But then here it's super far
from the prediction.

100
00:05:04,933 --> 00:05:06,033
And here as well.

101
00:05:06,033 --> 00:05:08,033
So clearly not well adapted.

102
00:05:08,033 --> 00:05:10,166
And that's why I wanted to show you this.

103
00:05:10,166 --> 00:05:11,566
Because now you're going to see

104
00:05:11,566 --> 00:05:15,733
that the polynomial regression
results are going to be much better.

105
00:05:15,900 --> 00:05:17,966
And now I'm going to show this to you
right away.

106
00:05:17,966 --> 00:05:19,400
So I'm going to close this.

107
00:05:19,400 --> 00:05:22,133
And we're going to visualize
the polynomial regression result

108
00:05:22,133 --> 00:05:25,133
efficiently by taking this code.

109
00:05:25,366 --> 00:05:28,200
Because you're going to see that it's
going to be almost the same.

110
00:05:28,200 --> 00:05:32,500
So I'm going to scroll down
I'll scroll back up to compare.

111
00:05:32,500 --> 00:05:36,366
Create a new code
cell facing the here and now.

112
00:05:36,366 --> 00:05:39,600
According to you,
what do we have to change here?

113
00:05:40,066 --> 00:05:41,700
Actually something important.

114
00:05:41,700 --> 00:05:46,233
Please press pause on the video now
and try to figure out before

115
00:05:46,233 --> 00:05:47,233
I give you the solution,

116
00:05:47,233 --> 00:05:50,533
what you have to replace here
in order to visualize the polynomial

117
00:05:50,533 --> 00:05:51,600
regression result.

118
00:05:51,600 --> 00:05:54,666
So please press pause in the video
and make that change in order

119
00:05:54,666 --> 00:05:56,900
to make it work
for polynomial regression results.