1
00:00:00,066 --> 00:00:02,100
Hello and welcome to this art tutorial.

2
00:00:02,100 --> 00:00:05,066
We are finally getting to the final round
of our classification

3
00:00:05,066 --> 00:00:07,000
adventure in our machine learning journey.

4
00:00:07,000 --> 00:00:11,633
Because today we're making the very last
classifier, the random Forest classifier.

5
00:00:12,000 --> 00:00:13,300
So let's start with the basics.

6
00:00:13,300 --> 00:00:14,700
We'll go to our Machine Learning A-Z

7
00:00:14,700 --> 00:00:17,533
folder to set the right
folder as working directory.

8
00:00:17,533 --> 00:00:20,600
So this right
folder is in by three classification.

9
00:00:20,933 --> 00:00:23,633
And here it is
section 20 random forest classification.

10
00:00:23,633 --> 00:00:25,300
That's the right folder.

11
00:00:25,300 --> 00:00:29,100
And as I was telling you
we are reaching the end of part

12
00:00:29,100 --> 00:00:33,266
three classification because the next
section will be dedicated to evaluating

13
00:00:33,266 --> 00:00:36,633
classification models performance
in order to improve the models.

14
00:00:37,100 --> 00:00:39,300
And therefore this is our last classifier.

15
00:00:39,300 --> 00:00:40,600
So let's go to this folder.

16
00:00:40,600 --> 00:00:42,600
That's the folder
we want to set its working directory.

17
00:00:42,600 --> 00:00:45,533
Make sure that you have the social network
add CSV file.

18
00:00:45,533 --> 00:00:46,466
And if that's the case

19
00:00:46,466 --> 00:00:49,966
click on this more button here
and then set as working directory.

20
00:00:50,466 --> 00:00:51,300
All good here.

21
00:00:51,300 --> 00:00:54,800
And now let's
go to our classification template to build

22
00:00:54,800 --> 00:00:57,833
our random forest classifier
with the best efficiency.

23
00:00:58,066 --> 00:01:00,900
So we'll just take everything from here

24
00:01:00,900 --> 00:01:03,400
to the bottom copy

25
00:01:03,400 --> 00:01:06,433
and then we'll paste it here.

26
00:01:06,966 --> 00:01:07,333
All right.

27
00:01:07,333 --> 00:01:10,333
So I'll just give a quick reminder
about what this template is doing.

28
00:01:10,633 --> 00:01:13,233
So the first step is to import
the data set.

29
00:01:13,233 --> 00:01:15,566
Then we encode here
the target feature as vector.

30
00:01:15,566 --> 00:01:19,333
Because we're just doing this to specify
to the classifier that our last column

31
00:01:19,333 --> 00:01:23,866
purchased is a categorical variable
with factor levels zero and one.

32
00:01:24,200 --> 00:01:26,566
So we don't need to do it
for all the classifiers.

33
00:01:26,566 --> 00:01:30,300
But we do need to do it for some,
like the Naive Bayes classifier as we saw.

34
00:01:30,600 --> 00:01:33,600
So it's good to keep this code section
in the template.

35
00:01:33,600 --> 00:01:36,833
Then we split the data set
into the training set and the test set.

36
00:01:36,833 --> 00:01:37,933
And then in the last section

37
00:01:37,933 --> 00:01:41,133
of the pre-processing phase
we apply features getting to the data.

38
00:01:41,133 --> 00:01:44,900
And we are only doing this because
at the end of this template we make this

39
00:01:44,900 --> 00:01:48,733
very cool graph for both the training
set result and the test set results

40
00:01:48,733 --> 00:01:52,066
that will plot the prediction regions
and the prediction boundary.

41
00:01:52,466 --> 00:01:55,466
Okay, so right now we need to change
a few things in this template.

42
00:01:55,533 --> 00:01:59,100
First, let's not forget to change
the title of the plot here

43
00:01:59,100 --> 00:02:02,100
will replace classifier by random

44
00:02:02,366 --> 00:02:05,366
forest classification.

45
00:02:05,500 --> 00:02:09,700
We'll copy this because we'll do the same
in the test set results here.

46
00:02:09,700 --> 00:02:12,066
It is perfect okay.

47
00:02:12,066 --> 00:02:16,600
And now what we only need to do
is to create our classifier here.

48
00:02:16,700 --> 00:02:18,333
And then we'll be ready.

49
00:02:18,333 --> 00:02:20,766
So I'll just space the same thing here.

50
00:02:20,766 --> 00:02:23,266
Random forest
classification to the training set.

51
00:02:23,266 --> 00:02:25,933
And now let's create our classifier.

52
00:02:25,933 --> 00:02:26,333
All right.

53
00:02:26,333 --> 00:02:28,500
Let's do this. So it's very simple.

54
00:02:28,500 --> 00:02:31,900
We're going to use a library
that is called random forest.

55
00:02:32,300 --> 00:02:33,000
So let's do this.

56
00:02:33,000 --> 00:02:37,800
Let's first install the package
with the command install that packages.

57
00:02:38,200 --> 00:02:42,300
So that's for those of you who are using
this package for the first time.

58
00:02:42,300 --> 00:02:44,933
Then you won't have
it installed on your packages.

59
00:02:44,933 --> 00:02:46,033
This here.

60
00:02:46,033 --> 00:02:49,200
As you can see, mine
should be already here because I used it

61
00:02:49,200 --> 00:02:49,700
several times.

62
00:02:49,700 --> 00:02:52,266
Yes, there it is, random forest.

63
00:02:52,266 --> 00:02:56,400
So if it's not the case for you
then you need to install the package here.

64
00:02:56,700 --> 00:02:59,433
And so in this install the packages

65
00:02:59,433 --> 00:03:03,066
function you just need to input the name
of the library in quotes.

66
00:03:03,400 --> 00:03:06,033
So it's random forest.

67
00:03:06,033 --> 00:03:10,033
Be careful with the capital F here
and not capital R here.

68
00:03:10,300 --> 00:03:12,700
All right.
So that's what you need to import anyway.

69
00:03:12,700 --> 00:03:14,200
That will install the package.

70
00:03:14,200 --> 00:03:16,800
I won't do it
because I already have mine installed.

71
00:03:16,800 --> 00:03:20,666
But what you just need to do is to select
this and press Command and Control.

72
00:03:20,666 --> 00:03:21,900
Plus enter to execute.

73
00:03:21,900 --> 00:03:24,900
And this will install the package
without any issues.

74
00:03:25,200 --> 00:03:25,500
All right.

75
00:03:25,500 --> 00:03:29,966
So here I'll just press command
plus shift plus C to make it as a comment.

76
00:03:30,200 --> 00:03:34,166
But then what I need to do
is to add the command here library

77
00:03:34,600 --> 00:03:37,600
and then input the name of the Random
Forest library.

78
00:03:37,833 --> 00:03:39,900
Because as you can see
here, it's not selected.

79
00:03:39,900 --> 00:03:44,900
So I need to add this library random
forest command to select it automatically,

80
00:03:44,900 --> 00:03:48,400
especially if I want to make
some automated scripts in the future.

81
00:03:48,966 --> 00:03:51,833
So that's the thing to do
that's very practical.

82
00:03:51,833 --> 00:03:53,166
And now we're all good.

83
00:03:53,166 --> 00:03:55,600
Let's create our classifier okay.

84
00:03:55,600 --> 00:03:58,600
So as usual we'll start
by creating the variable classifier.

85
00:03:59,233 --> 00:04:01,900
That will be a random forest
classifier itself.

86
00:04:01,900 --> 00:04:05,433
And now we will use the random forest
function to build our classifier.

87
00:04:05,966 --> 00:04:08,966
So here I'll just take the function
random forest.

88
00:04:09,733 --> 00:04:12,033
And now let's see what arguments
we need to input.

89
00:04:12,033 --> 00:04:17,566
So we'll just go here and press
F1 to get some info about this function.

90
00:04:17,566 --> 00:04:19,966
So here we need to click here.

91
00:04:19,966 --> 00:04:24,366
And here are some info about the Random
Forest library and function.

92
00:04:24,600 --> 00:04:27,600
So let's look at the arguments
which are which we are interested in.

93
00:04:28,200 --> 00:04:29,766
Okay. So the first argument is data.

94
00:04:29,766 --> 00:04:30,400
We will need it.

95
00:04:30,400 --> 00:04:33,033
As you can see
it's an optional data frame.

96
00:04:33,033 --> 00:04:36,900
So it's not an argument that we need
to build a random forest classifier.

97
00:04:37,333 --> 00:04:39,433
Same for subset and action.

98
00:04:39,433 --> 00:04:41,400
We actually don't need the three
first parameters

99
00:04:41,400 --> 00:04:42,933
to build a random forest classifier.

100
00:04:42,933 --> 00:04:48,400
However, we of course need the x and y
arguments here, which as you can

101
00:04:48,400 --> 00:04:52,500
guess, will be the matrix of features
and the dependent variable vector.

102
00:04:53,100 --> 00:04:57,433
Indeed, as you can see, x is a data frame
or a matrix of predictors.

103
00:04:57,433 --> 00:04:58,600
So that's pretty clear.

104
00:04:58,600 --> 00:05:01,966
X is our matrix of features
a matrix of predictors

105
00:05:01,966 --> 00:05:05,433
which are independent
variables age and estimated salary.

106
00:05:05,933 --> 00:05:08,566
And then y is set to be a response vector.

107
00:05:08,566 --> 00:05:11,400
So of course
y is the dependent variable vector

108
00:05:11,400 --> 00:05:14,933
that is the purchased column okay perfect.

109
00:05:14,933 --> 00:05:16,200
And then the last argument

110
00:05:16,200 --> 00:05:19,266
that we will need to build
our random forest classifier is

111
00:05:19,266 --> 00:05:22,533
of course the number of trees
we want to have in the forest.

112
00:05:23,066 --> 00:05:27,033
And this number of trees is given
by this entry argument here,

113
00:05:27,300 --> 00:05:30,600
which as you can see,
is the number of trees to grow.

114
00:05:30,933 --> 00:05:35,633
So to grow is a nice way of telling that
the trees will learn from the data set.

115
00:05:35,633 --> 00:05:37,133
How to make the predictions.

116
00:05:37,133 --> 00:05:40,100
And basically this argument
is the number of trees.

117
00:05:40,100 --> 00:05:43,000
So remember in the Python tutorial
we chose ten trees.

118
00:05:43,000 --> 00:05:44,500
Well let's do the same here.

119
00:05:44,500 --> 00:05:46,933
Let's pick N3 equals ten.

120
00:05:46,933 --> 00:05:52,000
And that is actually everything we need to
build our random forest classifier okay.

121
00:05:52,000 --> 00:05:52,733
So let's do this.

122
00:05:52,733 --> 00:05:54,066
Let's input the arguments.

123
00:05:54,066 --> 00:05:56,133
Let's go back to R and here we are.