1
00:00:00,766 --> 00:00:03,700
Okay, so now we imported the K-12 library.

2
00:00:03,700 --> 00:00:05,500
That is going to make a good split

3
00:00:05,500 --> 00:00:08,100
of the data set into the training set
and the test set.

4
00:00:08,100 --> 00:00:10,933
So now let's make the split.

5
00:00:10,933 --> 00:00:14,666
So remember in Python, for those of you
who followed the Python tutorial

6
00:00:14,800 --> 00:00:19,433
we used random state equals zero
so that we get the same results.

7
00:00:19,966 --> 00:00:21,700
Well here it's going to be the same.

8
00:00:21,700 --> 00:00:24,666
We're going to set
a seed to have the same results.

9
00:00:24,666 --> 00:00:26,900
Only we're going to do it
now. We're not going to wait

10
00:00:26,900 --> 00:00:30,100
to be in the function in the library
that makes this split.

11
00:00:30,366 --> 00:00:32,666
So to do it we have to set the same seed.

12
00:00:32,666 --> 00:00:35,900
And so we can type set dot seed.

13
00:00:36,200 --> 00:00:39,633
And in parenthesis
we can choose any number we want.

14
00:00:39,633 --> 00:00:41,200
It can be this number.

15
00:00:41,200 --> 00:00:42,900
You know that's the seed.

16
00:00:42,900 --> 00:00:45,700
If we both choose
this number will have the same results.

17
00:00:45,700 --> 00:00:47,366
But let's choose a more simple number.

18
00:00:47,366 --> 00:00:49,833
Let's just for example one two, three.

19
00:00:49,833 --> 00:00:52,100
And now let's make this split.

20
00:00:52,100 --> 00:00:55,100
It's not as simple as in Python
where we make it in one line.

21
00:00:55,433 --> 00:00:59,633
Here we're going to have to prepare the
method that we are going to call split.

22
00:01:00,100 --> 00:01:01,600
That is the method that is going to make

23
00:01:01,600 --> 00:01:05,500
the split of your data set into
the training set, and the test set equals

24
00:01:06,900 --> 00:01:09,100
and then sample

25
00:01:09,100 --> 00:01:12,100
dot split

26
00:01:12,333 --> 00:01:13,100
okay.

27
00:01:13,100 --> 00:01:16,300
Now if you want we can press F1
to see what we have to input.

28
00:01:16,433 --> 00:01:16,766
Okay.

29
00:01:16,766 --> 00:01:20,300
So it's simply split split data
into test and train set.

30
00:01:20,733 --> 00:01:21,633
Let's see the arguments.

31
00:01:21,633 --> 00:01:23,366
The first argument is why.

32
00:01:23,366 --> 00:01:25,966
So that's not the same argument
that we had to put for Python.

33
00:01:25,966 --> 00:01:26,400
In Python

34
00:01:26,400 --> 00:01:31,100
we had to put both the matrix of features
x and the dependent variable vector y.

35
00:01:31,200 --> 00:01:35,300
Here we only have to put y only we're
going to take y the following way.

36
00:01:35,300 --> 00:01:39,266
We're going to type data
set dollar sign purchased.

37
00:01:39,500 --> 00:01:42,366
Because your dependent variable
is purchase okay.

38
00:01:42,366 --> 00:01:44,366
So that's okay for the first parameter.

39
00:01:44,366 --> 00:01:46,033
And then what is the second parameter.

40
00:01:46,033 --> 00:01:48,366
It's split ratio okay.

41
00:01:48,366 --> 00:01:51,600
So split
ratio let's write split ratio here.

42
00:01:51,933 --> 00:01:54,600
And so split ratio is just a percentage

43
00:01:54,600 --> 00:01:57,900
of the observations
that you want to put in your training set.

44
00:01:58,366 --> 00:02:00,000
So let's be careful.

45
00:02:00,000 --> 00:02:03,000
In Python
we put the percentage for the test set.

46
00:02:03,200 --> 00:02:05,500
And here
we have to put it for the training set.

47
00:02:05,500 --> 00:02:09,100
So remember in Python
we chose 20% for the test set.

48
00:02:09,233 --> 00:02:14,000
So here logically for the training set
we will choose 0.8 okay.

49
00:02:14,000 --> 00:02:15,966
So what we'll do split return.

50
00:02:15,966 --> 00:02:20,366
So it will return true or false
for each of your observations.

51
00:02:20,533 --> 00:02:23,533
That means that each observation
will have either true or false.

52
00:02:23,700 --> 00:02:24,900
And this is going to be true

53
00:02:24,900 --> 00:02:27,900
if this observation was chosen
to go to the training set,

54
00:02:28,066 --> 00:02:31,233
and false if the observation
was chosen to go to the test set.

55
00:02:31,666 --> 00:02:33,466
So let's have a look. Let's select this

56
00:02:34,833 --> 00:02:37,833
command or Control plus enter to execute.

57
00:02:38,100 --> 00:02:40,500
And here it is. We have the split here.

58
00:02:40,500 --> 00:02:43,500
So now let's
go to the console and write split.

59
00:02:43,633 --> 00:02:45,833
All right. Enter.

60
00:02:45,833 --> 00:02:46,533
All right.

61
00:02:46,533 --> 00:02:48,866
You see that you have ten values.

62
00:02:48,866 --> 00:02:51,866
And so true means that the observation
goes to the training set.

63
00:02:51,900 --> 00:02:55,500
And false means that the observation
goes to the test set okay.

64
00:02:56,200 --> 00:02:59,566
So now what we have to do
is to create the training set

65
00:02:59,566 --> 00:03:01,000
and the test set separately.

66
00:03:01,000 --> 00:03:01,833
So we're going to do this.

67
00:03:01,833 --> 00:03:05,633
We'll type training set which is going
to be the name of our training set.

68
00:03:05,633 --> 00:03:07,133
Actually let's make it simple.

69
00:03:07,133 --> 00:03:10,700
So training set equals subset parenthesis.

70
00:03:10,966 --> 00:03:13,900
And here is the first argument
we put data set

71
00:03:13,900 --> 00:03:16,800
because it's the training
set is a subset of the data set.

72
00:03:17,766 --> 00:03:19,033
And here we will specify

73
00:03:19,033 --> 00:03:22,233
that we want the split equals equals true.

74
00:03:22,900 --> 00:03:24,500
And that's it okay.

75
00:03:24,500 --> 00:03:25,600
So that's it for the training set.

76
00:03:25,600 --> 00:03:28,600
Now let's copy this line.

77
00:03:29,400 --> 00:03:30,600
Paste it here.

78
00:03:30,600 --> 00:03:34,866
And here we're going to change training
set to test set.

79
00:03:35,400 --> 00:03:38,466
And of course we're going to change
split equals equals true

80
00:03:39,166 --> 00:03:41,566
to false.

81
00:03:41,566 --> 00:03:45,333
Because the test set are the observations
for which the split equals false.

82
00:03:45,900 --> 00:03:46,866
And now we're ready.

83
00:03:46,866 --> 00:03:48,866
We are ready to make the splits

84
00:03:48,866 --> 00:03:50,933
of the data set into the training set
and the test set.

85
00:03:50,933 --> 00:03:54,400
So let's execute these lines
command and control.

86
00:03:54,400 --> 00:03:56,400
Plus enter to execute.

87
00:03:56,400 --> 00:03:57,600
And here we go.

88
00:03:57,600 --> 00:03:59,633
Test set and training sets are created.

89
00:03:59,633 --> 00:04:01,000
Let's look at them.

90
00:04:01,000 --> 00:04:03,700
So let's click on the training set here.

91
00:04:03,700 --> 00:04:05,600
And the test set.

92
00:04:05,600 --> 00:04:08,600
Let's move that here and that here okay.

93
00:04:08,833 --> 00:04:10,600
Let's look at the training set first okay.

94
00:04:10,600 --> 00:04:13,600
We see that we have eight observations
okay good.

95
00:04:13,633 --> 00:04:15,800
And now let's look at the test set.

96
00:04:15,800 --> 00:04:20,333
Now we can clearly see that we have
two observations with six and nine perfect

97
00:04:22,100 --> 00:04:22,566
okay.

98
00:04:22,566 --> 00:04:24,800
So that's it for this tutorial.

99
00:04:24,800 --> 00:04:29,400
You now know how to split your data
set into a training set and a test set.

100
00:04:29,800 --> 00:04:32,233
This is a must do in any machine
learning model.

101
00:04:32,233 --> 00:04:33,433
You have to test

102
00:04:33,433 --> 00:04:37,133
the performance of your machine
learning model into a separate test set.

103
00:04:37,600 --> 00:04:38,766
So congratulations!

104
00:04:38,766 --> 00:04:39,900
Now you are almost ready

105
00:04:39,900 --> 00:04:44,366
to begin the journey of making exciting
machine learning models.

106
00:04:44,366 --> 00:04:47,366
We just have one thing to do left
feature scaling.

107
00:04:47,400 --> 00:04:50,400
You'll understand in the next tutorial
why it's so important to do this.

108
00:04:50,400 --> 00:04:51,966
So I look forward to seeing you there.

109
00:04:51,966 --> 00:04:54,966
And until then, enjoy machine learning.