1
00:00:00,100 --> 00:00:02,666
Hello and welcome to this art tutorial.

2
00:00:02,666 --> 00:00:07,033
Today we're going to implement
the k nearest neighbors classifier on R.

3
00:00:07,333 --> 00:00:08,666
So let's start right now.

4
00:00:08,666 --> 00:00:11,300
The first thing we're going to do
is to set the working directory.

5
00:00:11,300 --> 00:00:12,966
So right now I'm on my desktop.

6
00:00:12,966 --> 00:00:16,833
We're going to go to Machine Learning
A-Z folder part three classification.

7
00:00:17,100 --> 00:00:19,800
And then section K Nearest neighbors.

8
00:00:19,800 --> 00:00:22,333
All right. That's the folder
we want to set. It's working directory.

9
00:00:22,333 --> 00:00:25,533
Make sure that you have the social network
add CSV file.

10
00:00:25,933 --> 00:00:28,600
And if that's the case
you're ready to click on this more button

11
00:00:28,600 --> 00:00:31,600
here to set the folder as working
directory.

12
00:00:31,866 --> 00:00:33,300
Here we go. That's done.

13
00:00:33,300 --> 00:00:37,466
And now we're going to use our
classification template to be efficient

14
00:00:37,833 --> 00:00:41,100
and generate
all the results in a flashlight.

15
00:00:41,633 --> 00:00:45,700
So I'm going to 
select everything from here to the top.

16
00:00:46,133 --> 00:00:48,000
There you go. Copy and

17
00:00:49,200 --> 00:00:50,333
paste it here.

18
00:00:50,333 --> 00:00:52,700
And now
we need to change a very few things.

19
00:00:52,700 --> 00:00:55,200
First
we need to create the classifier here.

20
00:00:55,200 --> 00:00:57,833
And we just need to change the titles

21
00:00:57,833 --> 00:01:01,233
of our plots in the training set results
and the test results.

22
00:01:01,233 --> 00:01:04,233
So let's do this right now
so that we don't forget

23
00:01:04,266 --> 00:01:05,933
I'm going to replace classifier by.

24
00:01:05,933 --> 00:01:07,900
So it's k and ends.

25
00:01:07,900 --> 00:01:11,933
We're just doing this to specify
the algorithm that we're plotting.

26
00:01:12,800 --> 00:01:16,866
And here as well
classifier is replaced by k.

27
00:01:17,366 --> 00:01:19,066
And then right.

28
00:01:19,066 --> 00:01:23,000
And now what we only need to do
is to create our classifier.

29
00:01:23,433 --> 00:01:27,000
But first let's select
and execute the pre-processing steps.

30
00:01:27,000 --> 00:01:29,700
So I just need
to select everything from here

31
00:01:29,700 --> 00:01:32,700
and then press Command Control
plus enter to execute.

32
00:01:33,766 --> 00:01:35,100
That's done. All right.

33
00:01:35,100 --> 00:01:37,733
So we have our data sets.
We can have a look.

34
00:01:37,733 --> 00:01:41,866
That's our data
set our training set and our test set

35
00:01:42,433 --> 00:01:45,333
which are well scaled okay.

36
00:01:45,333 --> 00:01:48,633
So as a reminder
the data set contains information

37
00:01:48,766 --> 00:01:51,300
about users in a social network.

38
00:01:51,300 --> 00:01:54,600
The social network has a business client,
which is a car company

39
00:01:54,900 --> 00:01:57,733
that puts ads on the social network.

40
00:01:57,733 --> 00:02:01,366
And the social network gathered
not only the information

41
00:02:01,366 --> 00:02:04,500
like the age
and the estimated salary of those users,

42
00:02:04,766 --> 00:02:09,133
but it also took the response
to these users to the ad.

43
00:02:09,133 --> 00:02:10,766
That is zero.

44
00:02:10,766 --> 00:02:12,566
If the user didn't buy

45
00:02:12,566 --> 00:02:16,366
the product, the car, 
and one if the user bought the product.

46
00:02:16,800 --> 00:02:20,366
So, you know, it's a very cool luxury

47
00:02:20,366 --> 00:02:23,366
SUV launched at a ridiculously low price.

48
00:02:23,633 --> 00:02:28,433
So a lot of people, when they saw the ad
said, let's do this, let's get the car.

49
00:02:28,433 --> 00:02:30,533
And we can see there are a lot of buyers.

50
00:02:30,533 --> 00:02:33,933
So yeah, as you can see,
it must be a very cool car and cheap.

51
00:02:34,200 --> 00:02:39,000
So now let's go back to and
and create our classifier.

52
00:02:39,666 --> 00:02:41,366
So the classifier is a k.

53
00:02:41,366 --> 00:02:42,333
And then classifier.

54
00:02:42,333 --> 00:02:44,533
And to create this we're going to

55
00:02:45,566 --> 00:02:47,933
first import the right library.

56
00:02:47,933 --> 00:02:51,266
We need a library for this
which is by the way called class.

57
00:02:51,700 --> 00:02:54,700
So let's do this.

58
00:02:54,833 --> 00:02:57,300
So if by any chance the class

59
00:02:57,300 --> 00:03:01,733
library is not in your packages list,
you need to install it using this command.

60
00:03:01,733 --> 00:03:06,033
Install the packages and in parentheses
and in quotes the name of the class

61
00:03:06,033 --> 00:03:09,766
which is class, but
I think you might have it by default, so.

62
00:03:10,133 --> 00:03:13,933
But if that's not the case, please
write this command with class inside.

63
00:03:14,133 --> 00:03:16,633
Then execute and it will install it.

64
00:03:16,633 --> 00:03:19,633
But so here
we're just going to write library

65
00:03:19,966 --> 00:03:23,033
and in parentheses class
that's selected automatically.

66
00:03:23,033 --> 00:03:26,933
The class you see that right now
the class package is not selected.

67
00:03:27,300 --> 00:03:31,233
And once this is executed
it will select it automatically.

68
00:03:31,800 --> 00:03:32,666
All right.

69
00:03:32,666 --> 00:03:36,766
And now now we're going to do something
a little different

70
00:03:37,166 --> 00:03:39,733
because usually you know
we create a class.

71
00:03:39,733 --> 00:03:44,633
If I go here and then we create this
vector of prediction y using this predict

72
00:03:44,633 --> 00:03:49,333
function that we apply on our classifier
and on new observations in the test set.

73
00:03:49,800 --> 00:03:54,100
But here it's like we're going to do this
two steps all in once because

74
00:03:54,300 --> 00:03:57,300
we're actually going to remove this line.

75
00:03:57,366 --> 00:04:00,366
And we are actually going to replace this
command

76
00:04:00,366 --> 00:04:04,866
fitting K and end to the training set

77
00:04:06,233 --> 00:04:08,000
and predicting

78
00:04:08,000 --> 00:04:11,100
the test set results.

79
00:04:11,200 --> 00:04:16,200
Because all in once we're going to fit K
and into our training set and predict

80
00:04:16,200 --> 00:04:20,233
if the users of our test set by
yes or no, the SUV.

81
00:04:20,566 --> 00:04:21,600
So you're going to understand

82
00:04:21,600 --> 00:04:25,200
why we're going to directly
create a vector of prediction y pred here.

83
00:04:26,100 --> 00:04:29,133
And then,
you know, before we had classifier equals

84
00:04:29,366 --> 00:04:32,966
then the SVM or logistic regression

85
00:04:33,166 --> 00:04:37,033
here, it's going to be directly white
bread equals k and n

86
00:04:37,033 --> 00:04:40,300
because actually this k
and then function here will return

87
00:04:40,300 --> 00:04:44,133
the predictions
of the test set observations okay.

88
00:04:44,133 --> 00:04:46,700
So now let's have a look at this k
and in function.

89
00:04:46,700 --> 00:04:48,133
And you'll get it.

90
00:04:48,133 --> 00:04:51,233
So press F1 okay.

91
00:04:51,366 --> 00:04:53,966
So k nearest neighbor classification.

92
00:04:53,966 --> 00:04:55,800
Let's first input the arguments.

93
00:04:55,800 --> 00:04:58,300
The first argument is train.

94
00:04:58,300 --> 00:05:01,800
So you can guess that by train
they mean the training set.

95
00:05:02,200 --> 00:05:05,900
So here we need to specify
what the training set is.

96
00:05:06,100 --> 00:05:09,100
And actually since we call it training set
it's training set.

97
00:05:09,266 --> 00:05:10,466
So train set here.

98
00:05:11,766 --> 00:05:13,566
All right training set.

99
00:05:13,566 --> 00:05:16,566
However the training set
as you can see contains

100
00:05:17,066 --> 00:05:20,400
the independent variables
and the dependent variable.

101
00:05:20,666 --> 00:05:23,666
And we only need to take the independent
variables.

102
00:05:23,733 --> 00:05:28,100
We want to include the dependent variable
in the k in in function here.

103
00:05:28,100 --> 00:05:30,533
So here we need to take
only the first two columns.

104
00:05:30,533 --> 00:05:32,600
So we're going to add a bracket here.

105
00:05:32,600 --> 00:05:35,933
And then comma
because on the left of this comma

106
00:05:36,233 --> 00:05:39,933
all the lines that I'm taking all the
observations and that the right of this

107
00:05:39,933 --> 00:05:43,700
comma is the columns
I want to include in the training set.

108
00:05:44,133 --> 00:05:46,666
So it's all the columns
except for the last one.

109
00:05:46,666 --> 00:05:50,133
So I'm going to put here minus three

110
00:05:50,233 --> 00:05:53,833
which means I'm removing
the last column of my training set.