1
00:00:00,133 --> 00:00:01,166
So great.

2
00:00:01,166 --> 00:00:02,700
That's for the data set.

3
00:00:02,700 --> 00:00:05,300
And now we're going to move on
to the implementation.

4
00:00:05,300 --> 00:00:06,966
Simple linear regression

5
00:00:06,966 --> 00:00:10,600
IPython notebook which we're going to open
with Google Collaboratory.

6
00:00:10,800 --> 00:00:12,500
So that's simple linear regression.

7
00:00:12,500 --> 00:00:15,500
Let's
see how this implementation is structured.

8
00:00:15,733 --> 00:00:18,000
First we're going to import the libraries.

9
00:00:18,000 --> 00:00:19,766
You recognize the first steps of the data

10
00:00:19,766 --> 00:00:22,466
preprocessing phase
which we covered in part one.

11
00:00:22,466 --> 00:00:25,333
Second we're going to import the data set.

12
00:00:25,333 --> 00:00:28,566
Then third we're going to split that data
set into the training set.

13
00:00:28,566 --> 00:00:29,866
And test set.

14
00:00:29,866 --> 00:00:32,700
Then we're going to train
the simple linear regression

15
00:00:32,700 --> 00:00:34,233
model on the training set.

16
00:00:34,233 --> 00:00:36,566
Then we're going to predict
the test results.

17
00:00:36,566 --> 00:00:37,866
Then we're going to visualize the training

18
00:00:37,866 --> 00:00:40,066
set results
and visualize the test results.

19
00:00:40,066 --> 00:00:41,100
But I don't want to show you this

20
00:00:41,100 --> 00:00:44,566
too much right now because I want to keep
the surprise for the end.

21
00:00:44,933 --> 00:00:48,000
And so what we're going to do
now is implement

22
00:00:48,133 --> 00:00:51,666
this whole simple linear
regression model from scratch together.

23
00:00:52,333 --> 00:00:56,166
And to do so,
we are going to create a copy of this file

24
00:00:56,333 --> 00:00:59,833
by clicking here on save a copy in Drive.

25
00:01:00,100 --> 00:01:04,066
This will create a new copy on your drive,
in which you will be able

26
00:01:04,066 --> 00:01:07,400
to do some modifications or implement
the new model.

27
00:01:07,966 --> 00:01:09,766
Now we just have to do one thing.

28
00:01:09,766 --> 00:01:13,800
It is to remove all the code cells
because we were going to re-implement it

29
00:01:13,800 --> 00:01:16,133
from scratch step by step.
So let's do this.

30
00:01:16,133 --> 00:01:20,433
We just need to click this trash
button here on only the code cells.

31
00:01:20,433 --> 00:01:23,500
Make sure to keep the text
in order to highlight you know,

32
00:01:23,700 --> 00:01:26,700
the structure of this implementation.

33
00:01:26,800 --> 00:01:27,700
So there we go.

34
00:01:27,700 --> 00:01:30,700
Let's just delete everything here.

35
00:01:30,933 --> 00:01:33,066
And one last. Perfect.

36
00:01:33,066 --> 00:01:35,700
So now we have the whole structure
of the implementation.

37
00:01:35,700 --> 00:01:39,100
You can clearly see the different steps
that we're going to implement together.

38
00:01:39,366 --> 00:01:42,300
I will ask you to implement
some of them on your own.

39
00:01:42,300 --> 00:01:44,866
And of course then
we will implement the solution together.

40
00:01:44,866 --> 00:01:47,166
But that's because
I really want you to learn by doing.

41
00:01:47,166 --> 00:01:49,100
I really want you to take action

42
00:01:49,100 --> 00:01:52,100
and try to implement
some parts of the models on your own.

43
00:01:52,800 --> 00:01:53,533
All right. Perfect.

44
00:01:53,533 --> 00:01:57,366
So now I want you to think
what is going to be our first step here

45
00:01:57,833 --> 00:02:00,833
based on,
you know, what we covered in part one.

46
00:02:01,200 --> 00:02:04,500
Well, our first step is obviously the data

47
00:02:04,500 --> 00:02:08,166
preprocessing phase
in which we have to import the data set

48
00:02:08,400 --> 00:02:12,800
and maybe use some tools
in order to preprocess it the right way.

49
00:02:12,933 --> 00:02:14,733
So that our future simple linear

50
00:02:14,733 --> 00:02:18,200
regression model can be ready
to be trained on this data set.

51
00:02:18,833 --> 00:02:19,533
All right.

52
00:02:19,533 --> 00:02:22,966
And now I'm excited because I'm going
to show you the efficiency

53
00:02:23,200 --> 00:02:26,966
of the data preprocessing template
and how easy and fast

54
00:02:26,966 --> 00:02:30,266
it will be for us to preprocess
our data set here.

55
00:02:30,500 --> 00:02:34,133
So what we're going to do now
is we're going to go to that previous

56
00:02:34,133 --> 00:02:38,966
folder, you know, the data
preprocessing folder which was in part

57
00:02:39,300 --> 00:02:43,633
one data preprocessing
accessing the Python section.

58
00:02:43,900 --> 00:02:44,633
And there we go.

59
00:02:44,633 --> 00:02:47,633
Now we're going to open that template.

60
00:02:48,266 --> 00:02:49,566
Perfect.

61
00:02:49,566 --> 00:02:52,866
And you're going to see
we're simply going to do some copy paste.

62
00:02:53,200 --> 00:02:56,066
And you'll see that
we will only have one thing to change.

63
00:02:56,066 --> 00:02:59,666
And then the data preprocessing
phase will be ready and done.

64
00:02:59,700 --> 00:03:03,666
It will be done for us to move on
with the next step, which will be,

65
00:03:03,666 --> 00:03:07,700
of course, to train the simple linear
regression model on the training set.

66
00:03:07,966 --> 00:03:08,633
All right.

67
00:03:08,633 --> 00:03:10,866
So now I'm going to show you
what I've just said.

68
00:03:10,866 --> 00:03:13,866
And simply going to do some
copy paste here.

69
00:03:13,866 --> 00:03:17,000
There we go to import first the libraries

70
00:03:17,000 --> 00:03:20,233
and then adding them in a new code cell.

71
00:03:20,233 --> 00:03:26,766
Then to import the data set I'm simply
going to copy paste this second code cell.

72
00:03:26,766 --> 00:03:29,766
And you'll see that
I will have almost nothing to change.

73
00:03:30,033 --> 00:03:33,033
So creating a new code
cell here, pasting it.

74
00:03:33,133 --> 00:03:33,966
And finally,

75
00:03:35,033 --> 00:03:37,000
let's do that final

76
00:03:37,000 --> 00:03:40,700
step of the data preprocessing phase
to split the data

77
00:03:40,700 --> 00:03:45,166
set into the training set and the test
set to ensure I remind, have the separate

78
00:03:45,400 --> 00:03:49,200
entities where we're going to train
separately the model and then evaluate it.

79
00:03:49,500 --> 00:03:51,600
So there we go. Let's paste it.

80
00:03:51,600 --> 00:03:54,900
And now as I've told you,
we only have one little thing

81
00:03:54,900 --> 00:03:58,200
to change, which is of course
the name of the data set.

82
00:03:58,500 --> 00:04:01,466
And I remind the name of the data set is

83
00:04:01,466 --> 00:04:05,200
we're going to go back to our regression
folder here

84
00:04:05,600 --> 00:04:09,700
very quickly two regression
simple linear regression Python.

85
00:04:09,700 --> 00:04:11,100
And there we go. Here we are.

86
00:04:11,100 --> 00:04:14,166
And as you can see
this is salary underscore data.

87
00:04:14,200 --> 00:04:17,733
So the only thing I have to change
here is the name of the data set.

88
00:04:18,000 --> 00:04:20,333
And then voila everything is ready.

89
00:04:20,333 --> 00:04:22,600
The data preprocessing phase is ready.

90
00:04:22,600 --> 00:04:25,900
Because indeed this takes
automatically the features

91
00:04:25,900 --> 00:04:28,900
because it selects all the columns
except the last one.

92
00:04:28,933 --> 00:04:32,100
And this selects automatically
the dependent variable vector

93
00:04:32,366 --> 00:04:33,866
because it selects the last column.

94
00:04:33,866 --> 00:04:36,133
So here
of course we have only one feature,

95
00:04:36,133 --> 00:04:39,566
therefore one column for the features
and one column for the dependent variable.

96
00:04:39,733 --> 00:04:42,566
But you will see, for example,
that for multiple linear regression

97
00:04:42,566 --> 00:04:43,800
it will be exactly the same.

98
00:04:43,800 --> 00:04:46,633
We will have absolutely nothing
to change here.

99
00:04:46,633 --> 00:04:47,600
Okay.

100
00:04:47,600 --> 00:04:50,100
So now let me show you
the beauty of the result.

101
00:04:50,100 --> 00:04:53,433
Don't forget
of course to upload the data set before,

102
00:04:53,733 --> 00:04:55,733
you know running these cells here.

103
00:04:55,733 --> 00:04:59,966
So to upload the data set
you indeed need to click this folder here.

104
00:05:00,300 --> 00:05:03,600
Then it's going to connect to a runtime
to enable file browsing,

105
00:05:03,900 --> 00:05:07,100
after which you'll be able to upload
your data set.

106
00:05:07,133 --> 00:05:08,166
There we go.

107
00:05:08,166 --> 00:05:10,366
So you click this upload button here.

108
00:05:10,366 --> 00:05:14,333
Then you go to, you know,
this whole machinery is that folder

109
00:05:14,333 --> 00:05:17,600
which was provided to you
in just the previous tutorial.

110
00:05:17,800 --> 00:05:20,100
You know, which you had to download
if not already.

111
00:05:20,100 --> 00:05:22,400
So indeed,
we're going to go inside this folder.

112
00:05:22,400 --> 00:05:24,666
Then we're going to go to part
two regression,

113
00:05:24,666 --> 00:05:27,333
then simple linear regression,
then Python.

114
00:05:27,333 --> 00:05:28,266
And there we go.

115
00:05:28,266 --> 00:05:31,066
We select our data set salary data.

116
00:05:31,066 --> 00:05:34,866
We open it
and now it is indeed in Google Colab.

117
00:05:35,133 --> 00:05:36,200
So perfect.

118
00:05:36,200 --> 00:05:39,900
And now we're simply going to execute
each of these cells first

119
00:05:39,900 --> 00:05:44,333
importing the libraries
then importing the data set.

120
00:05:44,666 --> 00:05:45,233
Perfect.

121
00:05:45,233 --> 00:05:49,500
And now splitting the data set into
the training set and test set and done.

122
00:05:49,766 --> 00:05:54,300
Now the data preprocessing phase is over
and we can move on to the next step

123
00:05:54,466 --> 00:05:57,900
training the simple linear
regression model on the training set.