1
00:00:00,100 --> 00:00:00,966
Hello, my friends.

2
00:00:00,966 --> 00:00:03,066
All right, let's
see if you got this right.

3
00:00:03,066 --> 00:00:04,366
So what did we have to do?

4
00:00:04,366 --> 00:00:08,300
Well, first, the first natural step
was to go to our data preprocessing.

5
00:00:08,300 --> 00:00:09,966
Template. To grab all the.

6
00:00:09,966 --> 00:00:11,766
Essential tools, which are indeed.

7
00:00:11,766 --> 00:00:14,900
Well, the first steps of the data
preprocessing phase.

8
00:00:14,900 --> 00:00:15,900
So let's do this.

9
00:00:15,900 --> 00:00:18,900
That's
the first thing you were supposed to do.

10
00:00:19,133 --> 00:00:21,500
So I'm copying this cell.

11
00:00:21,500 --> 00:00:24,400
And then I'm. Basing it right here.

12
00:00:24,400 --> 00:00:24,900
In a new.

13
00:00:24,900 --> 00:00:27,633
Code cell for importing the libraries.

14
00:00:27,633 --> 00:00:29,600
There we go. First step done.

15
00:00:29,600 --> 00:00:31,800
Now to import. The data set.

16
00:00:31,800 --> 00:00:34,466
Well same. Let's go to our data
preprocessing.

17
00:00:34,466 --> 00:00:36,766
Template which indeed is meant.

18
00:00:36,766 --> 00:00:38,566
For us to be. Efficient.

19
00:00:38,566 --> 00:00:39,800
So I'm.

20
00:00:39,800 --> 00:00:41,033
Copying this. And then.

21
00:00:41,033 --> 00:00:43,266
We'll see what we have to replace.

22
00:00:43,266 --> 00:00:45,433
But once again
it is. Supposed to help you so.

23
00:00:45,433 --> 00:00:47,266
That you don't have to do much.

24
00:00:47,266 --> 00:00:49,233
All right. So just pasted it.

25
00:00:49,233 --> 00:00:50,433
And let's do you know.

26
00:00:50,433 --> 00:00:54,000
Just this for the last step of the data
preprocessing phase.

27
00:00:54,000 --> 00:00:55,366
You know, in any situation.

28
00:00:55,366 --> 00:00:57,000
Regardless of whether or not.

29
00:00:57,000 --> 00:00:58,566
Is needed, feature scaling.

30
00:00:58,566 --> 00:00:58,933
All right.

31
00:00:58,933 --> 00:01:02,400
So let's take that cell here. And.

32
00:01:02,766 --> 00:01:06,500
Let's copy that right here.

33
00:01:06,500 --> 00:01:08,533
In a new. Code cell.

34
00:01:08,533 --> 00:01:10,633
All right.
So that's the first essential steps.

35
00:01:10,633 --> 00:01:14,200
You know, which have been most of the time
when building a machine learning model.

36
00:01:14,300 --> 00:01:15,733
All right. So now next step.

37
00:01:15,733 --> 00:01:17,700
What do we have. To change inside.

38
00:01:17,700 --> 00:01:19,066
This data preprocessing phase.

39
00:01:19,066 --> 00:01:20,866
Well let's take it step by step.

40
00:01:20,866 --> 00:01:21,600
First step.

41
00:01:21,600 --> 00:01:23,833
Of course we.
Don't have anything to change.

42
00:01:23,833 --> 00:01:24,566
Now let's see.

43
00:01:24,566 --> 00:01:26,533
Second step. Well, there you go.

44
00:01:26,533 --> 00:01:28,533
Of course. Here
we have. To change the name of the.

45
00:01:28,533 --> 00:01:29,400
Data set.

46
00:01:29,400 --> 00:01:33,033
It is not data dot CSV, but it is now

47
00:01:33,033 --> 00:01:36,200
social network ads dot CSV.

48
00:01:36,200 --> 00:01:36,966
So let's do this.

49
00:01:36,966 --> 00:01:38,500
Let's just replace this.

50
00:01:38,500 --> 00:01:40,900
That was what you had to do next.

51
00:01:40,900 --> 00:01:43,833
Social network underscore.

52
00:01:43,833 --> 00:01:45,900
Ads dot CSV.

53
00:01:45,900 --> 00:01:47,533
And now next question.

54
00:01:47,533 --> 00:01:49,666
Do we have to change anything here.

55
00:01:49,666 --> 00:01:52,566
Well of course the answer is no.

56
00:01:52,566 --> 00:01:54,000
No, because. This data.

57
00:01:54,000 --> 00:01:56,000
Preprocessing template was meant.

58
00:01:56,000 --> 00:02:00,233
For. You to not have to change anything
in most situations, as.

59
00:02:00,233 --> 00:02:00,966
Long as of course,

60
00:02:00,966 --> 00:02:05,200
you make sure that in your data set
you have the features in the.

61
00:02:05,200 --> 00:02:07,866
First columns,
you know these are the. Two features here.

62
00:02:07,866 --> 00:02:10,466
And the dependent variable
in the last column.

63
00:02:10,466 --> 00:02:11,400
And since.

64
00:02:11,400 --> 00:02:13,766
This automatically selects all.

65
00:02:13,766 --> 00:02:14,966
The columns except.

66
00:02:14,966 --> 00:02:17,666
The last. One
and this automatically. Selects the.

67
00:02:17,666 --> 00:02:20,266
Last column in this.
Of course, regardless of.

68
00:02:20,266 --> 00:02:22,933
The number of. Features
you have in your data set.

69
00:02:22,933 --> 00:02:24,033
Well, there you go.

70
00:02:24,033 --> 00:02:26,366
This will simply select.

71
00:02:26,366 --> 00:02:27,300
The age and.

72
00:02:27,300 --> 00:02:29,866
Salary. In the matrix of. Features.

73
00:02:29,866 --> 00:02:32,966
And this line will simply select.

74
00:02:33,300 --> 00:02:33,833
Well.

75
00:02:33,833 --> 00:02:36,700
The dependent variable
in a nice 1D. Vector.

76
00:02:36,700 --> 00:02:38,100
All right so there you go.

77
00:02:38,100 --> 00:02:41,100
That was the reason
why I said it was so easy.

78
00:02:41,200 --> 00:02:44,266
You only had to change
the name of the dataset here.

79
00:02:44,266 --> 00:02:46,033
Social network.

80
00:02:46,033 --> 00:02:47,733
Ads dot CSV. Okay.

81
00:02:47,733 --> 00:02:49,900
And actually speaking of the dataset,
now that.

82
00:02:49,900 --> 00:02:51,566
We have the code to.

83
00:02:51,566 --> 00:02:53,700
Import. It, well let's not forget to.

84
00:02:53,700 --> 00:02:56,866
Upload it in a notebook
because in the same tutorial.

85
00:02:56,866 --> 00:02:59,866
We will actually. Run these. For cells.

86
00:02:59,933 --> 00:03:01,900
But before let's make sure to.

87
00:03:01,900 --> 00:03:03,900
Finish implementing them. So here.

88
00:03:03,900 --> 00:03:06,533
We are connecting to a runtime to enable.

89
00:03:06,533 --> 00:03:08,866
File. Browsing.
And in a second there you go.

90
00:03:08,866 --> 00:03:11,033
We should have the upload button.

91
00:03:11,033 --> 00:03:13,800
So now we're going to click. It
to find our.

92
00:03:13,800 --> 00:03:16,233
Whole machine learning dataset. Folder.

93
00:03:16,233 --> 00:03:17,800
And we're going to go inside.

94
00:03:17,800 --> 00:03:20,333
Then we're going to go to part
three classification.

95
00:03:20,333 --> 00:03:22,466
Then section 14. Logistic.

96
00:03:22,466 --> 00:03:24,266
Regression Python.

97
00:03:24,266 --> 00:03:27,333
And we're going to select our social.

98
00:03:27,333 --> 00:03:29,033
Network. Ads data set.

99
00:03:29,033 --> 00:03:32,433
It is executive same with of course
the H here the estimated.

100
00:03:32,433 --> 00:03:34,833
Salary and the purchase. Columns. There
we go.

101
00:03:34,833 --> 00:03:36,300
Let's click. Open.

102
00:03:36,300 --> 00:03:39,633
It is going to upload. It in our notebook.

103
00:03:39,900 --> 00:03:40,900
As we can see here.

104
00:03:40,900 --> 00:03:43,100
There you go. Perfect. We have it.

105
00:03:43,100 --> 00:03:45,433
So now we're ready to import it.
But first.

106
00:03:45,433 --> 00:03:47,166
I want to finish the implementation.

107
00:03:47,166 --> 00:03:48,800
Of this data preprocessing phase.

108
00:03:48,800 --> 00:03:50,933
So let's. Move on here to the next. Step.

109
00:03:50,933 --> 00:03:53,933
Splitting the data
set into the training set and test it.

110
00:03:54,000 --> 00:03:55,766
And here do we have anything to change.

111
00:03:55,766 --> 00:03:57,733
Well there you go again. Of course.

112
00:03:57,733 --> 00:03:58,800
No it is not.

113
00:03:58,800 --> 00:04:00,000
Compulsory to.

114
00:04:00,000 --> 00:04:01,800
Change anything here. This will.

115
00:04:01,800 --> 00:04:03,466
Automatically. Create your.

116
00:04:03,466 --> 00:04:04,666
Training set and a test.

117
00:04:04,666 --> 00:04:06,333
Set out of the whole.

118
00:04:06,333 --> 00:04:06,966
Data set.

119
00:04:06,966 --> 00:04:08,833
You know the whole data
set composed of the.

120
00:04:08,833 --> 00:04:10,233
Matrix of features X.

121
00:04:10,233 --> 00:04:12,133
And the dependent. Variable vector y.

122
00:04:12,133 --> 00:04:14,500
We're going to do some. Prints
right to show. You everything.

123
00:04:14,500 --> 00:04:16,433
Because then, you know the next.

124
00:04:16,433 --> 00:04:19,133
Step is to apply feature. Scaling.

125
00:04:19,133 --> 00:04:19,633
And before.

126
00:04:19,633 --> 00:04:22,566
We get to this well I just would like
to. Change something here.

127
00:04:22,566 --> 00:04:24,400
But again that was. Absolute.

128
00:04:24,400 --> 00:04:25,233
Not necessary.

129
00:04:25,233 --> 00:04:27,266
You could totally leave it like that.

130
00:04:27,266 --> 00:04:28,700
But I. Just want to change the.

131
00:04:28,700 --> 00:04:30,133
Test size because.

132
00:04:30,133 --> 00:04:30,533
Since.

133
00:04:30,533 --> 00:04:34,066
Actually we have 400 observations.

134
00:04:34,066 --> 00:04:34,800
In total, you know, we.

135
00:04:34,800 --> 00:04:37,800
Have 400. Customers in this data. Set.

136
00:04:38,000 --> 00:04:40,833
Well,
you know, if we choose a test size of.

137
00:04:40,833 --> 00:04:42,933
Oh point 25, it will be nice because.

138
00:04:42,933 --> 00:04:44,533
We will have actually three.

139
00:04:44,533 --> 00:04:48,766
Hundred customers in the training set
and 100 customers in the test.

140
00:04:48,766 --> 00:04:51,133
Set right will. Have nice round numbers.

141
00:04:51,133 --> 00:04:53,166
So just exceptionally here, I.

142
00:04:53,166 --> 00:04:55,300
Usually recommend. Indeed oh. Point two.

143
00:04:55,300 --> 00:04:57,733
As a test size, but oh. Point 25 is.

144
00:04:57,733 --> 00:04:59,133
Actually totally fine.

145
00:04:59,133 --> 00:05:03,433
We will have a nice training
set of 300 customers and one nice test.

146
00:05:03,433 --> 00:05:05,366
Set. Of 100 customers.

147
00:05:05,366 --> 00:05:06,200
So there. You go.

148
00:05:06,200 --> 00:05:08,466
Now we can move on to the next step.

149
00:05:08,466 --> 00:05:11,366
Feature scaling. Okay.

150
00:05:11,366 --> 00:05:11,766
All right.

151
00:05:11,766 --> 00:05:12,466
So how are.

152
00:05:12,466 --> 00:05:15,000
We going to implement
feature scaling. Here.

153
00:05:15,000 --> 00:05:16,866
Well there is nothing more simple.

154
00:05:16,866 --> 00:05:20,333
Thanks to our data preprocessing toolkit
which I've put here.

155
00:05:20,500 --> 00:05:21,466
And if you haven't.

156
00:05:21,466 --> 00:05:22,400
Well grab it.

157
00:05:22,400 --> 00:05:24,733
In part one data preprocessing.

158
00:05:24,733 --> 00:05:27,566
And now in our toolkit
we're going to scroll down

159
00:05:27,566 --> 00:05:31,033
to the bottom
to find that feature scaling tool, which.

160
00:05:31,033 --> 00:05:34,033
Is actually the. Last one right. Here.

161
00:05:34,033 --> 00:05:35,733
It is that feature scaling.

162
00:05:35,733 --> 00:05:38,266
So here we go. Let's grab it.

163
00:05:38,266 --> 00:05:40,066
You know
you just have to grab the first cell.

164
00:05:40,066 --> 00:05:41,366
Implementing the tool.

165
00:05:41,366 --> 00:05:43,500
Now let's.
Go back to our logistic regression.

166
00:05:43,500 --> 00:05:44,700
Implementation.

167
00:05:44,700 --> 00:05:47,766
Let's create
a new cut cell here for feature scaling.

168
00:05:47,933 --> 00:05:50,866
And let's. Paste it. Right inside.