﻿1
00:00:00,100 --> 00:00:02,166
Hello and welcome to this art tutorial.

2
00:00:02,166 --> 00:00:05,133
So in this tutorial, we're not going
to make any machine learning model.

3
00:00:05,133 --> 00:00:07,400
We're just going to make a template.

4
00:00:07,400 --> 00:00:09,766
And this will be our regression template.

5
00:00:09,766 --> 00:00:12,200
So it's a template
that is going to be very useful

6
00:00:12,200 --> 00:00:15,233
for the next non-linear regression models
we are going to make.

7
00:00:15,466 --> 00:00:19,000
And also you can use this template
to create a regression

8
00:00:19,000 --> 00:00:21,233
model for your specific problem.

9
00:00:21,233 --> 00:00:24,266
Because as you can see,
you will only have a few things to change

10
00:00:24,433 --> 00:00:27,433
so that you will build your machine
learning model very efficiently.

11
00:00:28,100 --> 00:00:31,166
So let's make this template
and to make it we're going to go

12
00:00:31,166 --> 00:00:35,800
from what we did in polynomial regression
and generalized the whole thing.

13
00:00:36,000 --> 00:00:40,366
So I'm going to take everything from here
up to the end copy.

14
00:00:40,866 --> 00:00:43,166
And I'm going to paste that
in the template.

15
00:00:43,166 --> 00:00:46,666
And now let's make this template okay.

16
00:00:46,666 --> 00:00:48,966
So first of all let's start
with the beginning.

17
00:00:48,966 --> 00:00:51,633
So the beginning
was about pre-processing the data.

18
00:00:51,633 --> 00:00:53,633
So actually
there is nothing to change here.

19
00:00:53,633 --> 00:00:57,000
We will keep of course these two lines
because we will work on the position

20
00:00:57,000 --> 00:01:00,566
salary CSV file and build our non-linear
regression models on it.

21
00:01:00,766 --> 00:01:03,600
And for the same reason as before,
we will need to split the data

22
00:01:03,600 --> 00:01:07,266
set into a training set and a test set,
because this is a small data set

23
00:01:07,266 --> 00:01:10,433
and we will leave
feature scaling as comments here, since

24
00:01:10,433 --> 00:01:14,733
most of the libraries don't require us
to apply feature scaling manually.

25
00:01:15,366 --> 00:01:17,866
So all good for the pre-processing phase.

26
00:01:17,866 --> 00:01:21,400
And this will be the first part
of our regression template.

27
00:01:21,900 --> 00:01:24,000
Great. Now let's move on to the next part.

28
00:01:24,000 --> 00:01:27,000
So the next part
is about creating the model,

29
00:01:27,266 --> 00:01:30,166
creating the model,
and fitting it to the data set.

30
00:01:30,166 --> 00:01:34,000
So first of all, in the previous tutorial
we built two regression models

31
00:01:34,400 --> 00:01:37,433
the linear regression model
and the polynomial regression model.

32
00:01:37,433 --> 00:01:38,033
And that was just

33
00:01:38,033 --> 00:01:41,933
to compare the two results and to show you
that the polynomial regression model

34
00:01:41,933 --> 00:01:46,533
was much more appropriate to our problem
because it was non-linear regression

35
00:01:46,533 --> 00:01:50,400
model as opposed to the linear
regression model, which is a linear model.

36
00:01:50,666 --> 00:01:54,266
So we only did that
so that you can understand very clearly

37
00:01:54,266 --> 00:01:57,966
the distinction between linear models
and non-linear models.

38
00:01:58,200 --> 00:02:00,566
But now that it's
getting pretty clear in our mind,

39
00:02:00,566 --> 00:02:04,000
we are going to get rid of this section
here about linear regression.

40
00:02:04,266 --> 00:02:07,200
Because now we get the idea
and we will only focus

41
00:02:07,200 --> 00:02:10,433
on the new non-linear regression
models themselves

42
00:02:10,533 --> 00:02:13,300
without comparing them to the linear
regression model.

43
00:02:13,300 --> 00:02:16,333
But to keep the graphic results
of the linear regression model in mind,

44
00:02:16,566 --> 00:02:20,433
keep in mind that it was a straight line,
not fitting very well to data set,

45
00:02:20,733 --> 00:02:24,100
and therefore was not much appropriate
to our data set.

46
00:02:24,100 --> 00:02:27,100
Where
the data is non-linearly distributed.

47
00:02:27,200 --> 00:02:30,700
Okay,
so we're going to get rid of this done.

48
00:02:31,433 --> 00:02:35,366
Now we only have this code section here
that creates the non-linear regression

49
00:02:35,366 --> 00:02:36,533
model itself.

50
00:02:36,533 --> 00:02:38,066
So we're going to get rid of this

51
00:02:38,066 --> 00:02:41,066
because this corresponds
to the polynomial regression.

52
00:02:41,233 --> 00:02:44,833
And to generalize
we will just add a come here

53
00:02:44,966 --> 00:02:49,633
to specify
that this is where we have to create our

54
00:02:50,933 --> 00:02:54,066
regressor right here okay.

55
00:02:54,600 --> 00:02:57,433
And here we can also replace
polynomial regression

56
00:02:57,433 --> 00:03:00,433
by the regression model.

57
00:03:01,333 --> 00:03:04,133
And we are all good with this section
okay.

58
00:03:04,133 --> 00:03:05,500
Now let's move on to the next one.

59
00:03:05,500 --> 00:03:08,400
The next one is about
visualizing the results.

60
00:03:08,400 --> 00:03:12,133
So for the same reason
as what we did in this section,

61
00:03:12,133 --> 00:03:14,400
to fit the regression model
to the data set,

62
00:03:14,400 --> 00:03:18,933
we will get rid of the graphic results
of the linear regression model

63
00:03:19,100 --> 00:03:20,400
because we have them in mind,

64
00:03:20,400 --> 00:03:24,366
and we only want to focus on our future
non-linear regression models.

65
00:03:24,566 --> 00:03:26,666
So let's get rid of this.

66
00:03:26,666 --> 00:03:29,466
And now we only have the graphic results

67
00:03:29,466 --> 00:03:32,533
of our future non-linear regression
model okay.

68
00:03:32,800 --> 00:03:34,400
And the great thing about this template

69
00:03:34,400 --> 00:03:37,200
is that in this section
we won't need to change anything.

70
00:03:37,200 --> 00:03:40,200
Well, almost anything
because we're going to replace

71
00:03:40,200 --> 00:03:43,133
polynomial regression
by regression model.

72
00:03:45,466 --> 00:03:46,800
And we are going to change

73
00:03:46,800 --> 00:03:51,266
this poly rag regressor here by regressor.

74
00:03:51,266 --> 00:03:55,066
Because you know before we called it
poly rag because we also had our linear

75
00:03:55,066 --> 00:03:56,733
regressor linear.

76
00:03:56,733 --> 00:03:58,766
And so we needed to make the distinction.

77
00:03:58,766 --> 00:04:02,100
But now since we got rid of the linear
regression model,

78
00:04:02,166 --> 00:04:06,200
we will call our future non-linear
regression model regressor here.

79
00:04:06,300 --> 00:04:09,500
And that's why we input two regressor here
and the predict function.

80
00:04:09,900 --> 00:04:13,300
So basically when we get to this section,
when we build our future

81
00:04:13,300 --> 00:04:14,900
non-linear regression model,

82
00:04:14,900 --> 00:04:19,500
the only thing that we will need to change
is actually the title here.

83
00:04:19,500 --> 00:04:21,666
Basically
we don't have anything to change here.

84
00:04:21,666 --> 00:04:24,666
And that's the great thing about this
template.

85
00:04:24,866 --> 00:04:27,600
However, of course,
if you're using this template on your data

86
00:04:27,600 --> 00:04:31,100
set in your problem, well, the only thing
that you will need to change

87
00:04:31,333 --> 00:04:32,333
are the names here.

88
00:04:32,333 --> 00:04:36,600
The independent variable, and also
the name of the dependent variable here.

89
00:04:36,866 --> 00:04:40,233
That is, you'll have to replace salary
by the name of your dependent variable.

90
00:04:40,633 --> 00:04:41,866
But that's all okay.

91
00:04:41,866 --> 00:04:45,933
And we are finally getting
to the last section of this code template,

92
00:04:46,400 --> 00:04:50,766
which is about predicting the final result
to get our final verdict.

93
00:04:51,233 --> 00:04:53,300
Okay, so same reason as before.

94
00:04:53,300 --> 00:04:56,433
We're going to get rid of the linear
regression part for this.

95
00:04:56,833 --> 00:04:57,866
So here we go.

96
00:04:57,866 --> 00:04:58,833
Good bye.

97
00:04:58,833 --> 00:05:02,700
And now we are only left
with the polynomial regression prediction

98
00:05:03,233 --> 00:05:04,466
okay great.

99
00:05:04,466 --> 00:05:06,800
And so now we just need
to change something here.

100
00:05:06,800 --> 00:05:09,600
Well first it's the name of the regressor.

101
00:05:09,600 --> 00:05:14,466
Because before the name of our polynomial
regression regressor was called poly rag.

102
00:05:14,700 --> 00:05:16,633
And now
since we won't make the distinction

103
00:05:16,633 --> 00:05:19,966
between a linear regression model
and our non-linear regression model,

104
00:05:20,133 --> 00:05:23,133
well,
our regressor will be called regressor.

105
00:05:23,333 --> 00:05:28,600
So here regressor and here you know
for polynomial regression we need to input

106
00:05:28,700 --> 00:05:33,666
the polynomial features of our data frame
that we created in our data frame.

107
00:05:33,900 --> 00:05:38,066
Well we won't need to do this
for the next non-linear models

108
00:05:38,066 --> 00:05:42,100
because this is strictly specific
to the polynomial regression model

109
00:05:42,100 --> 00:05:45,100
where you know
we need to add those polynomial features.

110
00:05:45,200 --> 00:05:49,300
And so we will get rid of this
and we'll get something very simple.

111
00:05:49,766 --> 00:05:53,966
It's actually this line
that will perfectly predict

112
00:05:53,966 --> 00:05:57,233
the salary of this 6.5 position level.
