﻿1
00:00:00,133 --> 00:00:02,366
Hello and welcome to this art tutorial.

2
00:00:02,366 --> 00:00:06,633
Today we are going to implement
the Naive Bayes algorithm on R.

3
00:00:06,900 --> 00:00:07,800
So let's start right now.

4
00:00:07,800 --> 00:00:10,800
Let's first set
the folder as working directory.

5
00:00:10,866 --> 00:00:14,166
So I'm going to part three classification.

6
00:00:14,600 --> 00:00:16,100
And that's the folder.

7
00:00:16,100 --> 00:00:18,800
So make sure you have the social network
add CSV file.

8
00:00:18,800 --> 00:00:19,933
And if that's the case

9
00:00:19,933 --> 00:00:23,400
click on this more button here
to set this folder as working directory.

10
00:00:24,000 --> 00:00:27,766
And now let's start implementing
Naive Bayes on R.

11
00:00:28,200 --> 00:00:30,733
So we're going to take our classification
template.

12
00:00:30,733 --> 00:00:34,366
We're going to take everything
from here to here.

13
00:00:36,766 --> 00:00:38,466
And then

14
00:00:38,466 --> 00:00:40,300
paste right here.

15
00:00:40,300 --> 00:00:43,033
And now
we just need to change a few things.

16
00:00:43,033 --> 00:00:47,166
So first let's change the titles r usual
so that we don't forget.

17
00:00:47,233 --> 00:00:50,400
So here we are
with the Naive Bayes classifier.

18
00:00:51,500 --> 00:00:54,400
Same for the training set.

19
00:00:54,400 --> 00:00:57,133
So that is plotting
the title in our graph.

20
00:00:57,133 --> 00:00:59,700
And now what we only need to do

21
00:00:59,700 --> 00:01:03,666
is to create our Naive Bayes classifier.

22
00:01:04,366 --> 00:01:06,000
So let's do this.

23
00:01:06,000 --> 00:01:08,333
Going to remove this.

24
00:01:08,333 --> 00:01:09,833
And now what are we going to do.

25
00:01:09,833 --> 00:01:10,633
Well it's actually funny

26
00:01:10,633 --> 00:01:14,300
because we are going to use a library
that we already used before.

27
00:01:14,400 --> 00:01:17,400
But for a different classifier

28
00:01:17,400 --> 00:01:20,400
it's the E 1071 library.

29
00:01:20,566 --> 00:01:23,200
So there is no surprise
why this library is very popular.

30
00:01:23,200 --> 00:01:27,133
One of the most popular ones,
because it contains a lot of tools

31
00:01:27,133 --> 00:01:30,900
tools for SVM and tools for Naive Bayes.

32
00:01:31,366 --> 00:01:33,633
So here again we'll use this library.

33
00:01:33,633 --> 00:01:37,066
So for those of you
who didn't follow the SVM tutorials on R,

34
00:01:37,300 --> 00:01:41,033
and for those of you who are starting R
for the first time, maybe you don't have

35
00:01:41,033 --> 00:01:45,600
the E 1071 packages
in your packages list.

36
00:01:45,900 --> 00:01:49,666
So if that's the case,
you just need to type the command here.

37
00:01:49,666 --> 00:01:54,433
Install dot packages
and then in quotes here in the parenthesis

38
00:01:54,733 --> 00:01:59,866
you just add the name of the library
which is E 1071.

39
00:02:00,533 --> 00:02:04,533
And then you just need to select this
and execute to install the package.

40
00:02:04,533 --> 00:02:07,533
I won't do it here
because I already have mine installed.

41
00:02:07,600 --> 00:02:10,300
But I promise you it will work properly.

42
00:02:10,300 --> 00:02:13,300
So I'm just going to put that in comment.

43
00:02:13,533 --> 00:02:14,133
Here we go.

44
00:02:14,133 --> 00:02:18,400
And now let's just also import
the library that is to select it

45
00:02:18,666 --> 00:02:19,633
in the packages list.

46
00:02:19,633 --> 00:02:21,033
Because here it's selected.

47
00:02:21,033 --> 00:02:22,700
But it might not always be the case.

48
00:02:22,700 --> 00:02:25,933
And if you want to automate your machine
learning scripts, it's better

49
00:02:25,933 --> 00:02:30,200
to always have this script line here
that selects automatically your packages

50
00:02:30,733 --> 00:02:33,333
through the library. E10 71.

51
00:02:33,333 --> 00:02:36,900
And now we are ready to start
creating our Naive Bayes classifier.

52
00:02:37,666 --> 00:02:39,133
Okay, so let's do this.

53
00:02:39,133 --> 00:02:41,900
As usual,
we will create a variable classifier.

54
00:02:44,466 --> 00:02:45,966
And then we will use the

55
00:02:45,966 --> 00:02:50,500
Naive Bayes
function of the E 1071 library.

56
00:02:50,833 --> 00:02:53,400
The E 1071 library contains
a lot of functions,

57
00:02:53,400 --> 00:02:57,000
and one of them is to create a Naive Bayes
classifier.

58
00:02:57,700 --> 00:03:00,800
So okay, so careful with the, capitals.

59
00:03:00,800 --> 00:03:04,933
So it's, not capital
N, the capital B, all right.

60
00:03:04,933 --> 00:03:07,133
But then, you know, 
R is helping you with the term.

61
00:03:07,133 --> 00:03:10,500
So you would have this,
Naive Bayes popping up here.

62
00:03:10,500 --> 00:03:12,666
So you just have to press enter.

63
00:03:12,666 --> 00:03:14,733
And now let's input the parameters.

64
00:03:14,733 --> 00:03:16,333
So the parameters what are they.

65
00:03:16,333 --> 00:03:17,400
Let's look at them.

66
00:03:17,400 --> 00:03:20,400
We press F1 here to have a look.

67
00:03:20,600 --> 00:03:22,233
And here there are.

68
00:03:22,233 --> 00:03:26,666
So we will actually need to input
only the two first arguments.

69
00:03:26,900 --> 00:03:30,133
We will need to import the formula
like for the other one.

70
00:03:30,133 --> 00:03:33,133
It will work perfectly
well with only the first two ones.

71
00:03:33,566 --> 00:03:34,266
So you'll see.

72
00:03:34,266 --> 00:03:38,533
Let's input x here,
which is by the way the training set.

73
00:03:39,300 --> 00:03:40,200
So the training set.

74
00:03:40,200 --> 00:03:41,933
But we need to remove

75
00:03:41,933 --> 00:03:45,533
the last column of our training set
because X is actually the matrix features.

76
00:03:45,533 --> 00:03:49,233
You see it's written here
in numeric matrix or data frame.

77
00:03:49,566 --> 00:03:53,833
And by that they mean
the matrix of features that is the matrix

78
00:03:54,133 --> 00:03:55,700
of the independent variable.

79
00:03:55,700 --> 00:03:59,466
So here, since the training set contains
both the independent variables

80
00:03:59,466 --> 00:04:03,033
and the dependent variable,
we need to exclude the dependent variable.

81
00:04:03,333 --> 00:04:06,200
So to do this
we're going to add brackets here

82
00:04:06,200 --> 00:04:10,466
and remove the last column
by inputting here a minus.

83
00:04:10,466 --> 00:04:14,100
And then the index of the last column
which is well we can see here

84
00:04:14,100 --> 00:04:17,100
it's minus three. So it's minus three.

85
00:04:17,233 --> 00:04:19,366
All right. So perfect.

86
00:04:19,366 --> 00:04:22,966
Then comma and then next argument.

87
00:04:23,100 --> 00:04:26,100
So do you guess what the next argument is.

88
00:04:26,100 --> 00:04:29,500
Well of course to train a classifier
we need the independent variables

89
00:04:29,500 --> 00:04:30,733
and the response.

90
00:04:30,733 --> 00:04:32,700
And the response
is the dependent variable.

91
00:04:32,700 --> 00:04:35,700
So of course
y is going to be the dependent variable.

92
00:04:36,000 --> 00:04:40,466
So y equals and to take it
we are going to take it this way.

93
00:04:40,466 --> 00:04:42,600
We're going to input training set

94
00:04:44,466 --> 00:04:47,333
dollar purchased.

95
00:04:47,333 --> 00:04:50,866
So I'm choosing to write it this way
because we can clearly see that

96
00:04:51,000 --> 00:04:54,000
we are taking the dependent
variable purchased.
