﻿1
00:00:00,133 --> 00:00:02,533
Hello and welcome to this art tutorial.

2
00:00:02,533 --> 00:00:05,100
Here we are
at the final round of regression

3
00:00:05,100 --> 00:00:08,433
with our final regression
model random Forest Regression.

4
00:00:09,000 --> 00:00:12,066
In the previous section
we saw the decision tree regression model.

5
00:00:12,333 --> 00:00:15,533
So if now the decision tree regression
doesn't have any secret for you,

6
00:00:15,600 --> 00:00:18,600
then you will perfectly understand
random forest regression

7
00:00:18,766 --> 00:00:22,100
because random forest is just a
team of decision trees,

8
00:00:22,400 --> 00:00:25,733
each one making some prediction
of your dependent variable

9
00:00:25,933 --> 00:00:29,466
and the ultimate prediction of the random
forest itself is simply

10
00:00:29,466 --> 00:00:33,266
the average of the different predictions
of all the different trees in the forest.

11
00:00:33,466 --> 00:00:35,233
And actually,
at the end of the previous section

12
00:00:35,233 --> 00:00:37,833
about decision trees,
I asked you an enigma.

13
00:00:37,833 --> 00:00:43,000
The enigma was knowing the results we got
with one tree, what would be the result?

14
00:00:43,000 --> 00:00:46,000
With ten trees or 100 trees or 500 trees

15
00:00:46,200 --> 00:00:49,400
in terms of visualization
and in term of prediction.

16
00:00:49,833 --> 00:00:53,966
So I hope that after watching
the intuition tutorial made by Kirill,

17
00:00:54,166 --> 00:00:57,600
you actually asked yourself this question
and tried to predict

18
00:00:57,600 --> 00:01:00,600
what's going to happen here
with random forest regression.

19
00:01:00,900 --> 00:01:02,366
So let's find out about that.

20
00:01:02,366 --> 00:01:06,333
We are going to build a random forest
regression model and see what happens.

21
00:01:06,533 --> 00:01:07,700
So let's do it.

22
00:01:07,700 --> 00:01:11,200
We're going to start by selecting
the right folder as a working directory.

23
00:01:11,200 --> 00:01:13,733
So it's in part two regression.

24
00:01:13,733 --> 00:01:15,466
And here is the final regression model.

25
00:01:15,466 --> 00:01:17,400
We were building random forest regression.

26
00:01:17,400 --> 00:01:18,733
So let's go inside.

27
00:01:18,733 --> 00:01:20,266
And that's the right folder we want to set

28
00:01:20,266 --> 00:01:23,266
is working directory with the position
salary CSV file.

29
00:01:23,400 --> 00:01:27,600
So let's click on this more button
and set as working directory.

30
00:01:27,966 --> 00:01:28,933
All good.

31
00:01:28,933 --> 00:01:32,733
And now let's take our regression template
to build this model efficiently.

32
00:01:33,000 --> 00:01:37,033
So we're actually going to take everything
from here to the bottom.

33
00:01:37,300 --> 00:01:40,533
But we will only include this code section
to visualize the regression

34
00:01:40,533 --> 00:01:41,533
model results.

35
00:01:41,533 --> 00:01:44,333
Because you understood
that the decision tree regression

36
00:01:44,333 --> 00:01:47,100
model is a non continuous regression
model.

37
00:01:47,100 --> 00:01:50,100
And since random
forest is a combination of decision trees

38
00:01:50,300 --> 00:01:53,133
then it's a combination of non
continuous regression model.

39
00:01:53,133 --> 00:01:56,766
And intuitively we understand
we can guess that the random forest

40
00:01:56,766 --> 00:01:59,766
regression model
is not going to be continuous either.

41
00:01:59,900 --> 00:02:03,866
So since this code doesn't work for non
continuous regression model

42
00:02:04,133 --> 00:02:05,666
we will actually use this one.

43
00:02:05,666 --> 00:02:07,533
That works perfectly for it.

44
00:02:07,533 --> 00:02:12,300
So I'm going to copy this paste
that here and remove this section.

45
00:02:12,300 --> 00:02:16,100
That is non appropriate for non continuous
regression models.

46
00:02:16,333 --> 00:02:17,133
Here we go.

47
00:02:17,133 --> 00:02:19,500
And now the template is ready.

48
00:02:19,500 --> 00:02:21,066
Let's change the basics.

49
00:02:21,066 --> 00:02:24,266
Let's replace here
regression model by random forest

50
00:02:25,200 --> 00:02:28,066
regression.

51
00:02:28,066 --> 00:02:30,966
Visualizing the random forest
regression results

52
00:02:30,966 --> 00:02:36,000
and fitting random forest
regression to our dataset okay great.

53
00:02:36,200 --> 00:02:39,833
So now let's build the model
which is in this section here.

54
00:02:39,833 --> 00:02:42,600
So let's remove this.

55
00:02:42,600 --> 00:02:46,666
And as usual we're going to import
the right library for the job.

56
00:02:46,866 --> 00:02:51,500
And then use a function
to build our random forest regressor.

57
00:02:51,966 --> 00:02:55,400
So the package we are going to import
is called random forest.

58
00:02:55,833 --> 00:02:58,833
So for those of you
who don't have the package installed

59
00:02:59,100 --> 00:03:02,100
in your packages here,
well you can check it out.

60
00:03:02,100 --> 00:03:05,100
Mine is already installed
because I used it before,

61
00:03:05,100 --> 00:03:09,933
but I'm going to write this line here
for those of you who need to install it.

62
00:03:10,200 --> 00:03:13,200
So install dot packages,

63
00:03:13,500 --> 00:03:16,900
parenthesis and in quotes random.

64
00:03:17,633 --> 00:03:20,400
So no capital R but then capital F.

65
00:03:20,400 --> 00:03:23,133
Oh right. All right random forest.

66
00:03:23,133 --> 00:03:25,766
And so I'm not going to install it
because mine is already installed.

67
00:03:25,766 --> 00:03:27,266
So I'm going to put that in comment.

68
00:03:27,266 --> 00:03:30,100
But if you want to install it
you just need to select this line as

69
00:03:30,100 --> 00:03:33,133
I just did and press Command and Control
plus enter to execute it.

70
00:03:33,466 --> 00:03:35,800
And this will install the package
properly.

71
00:03:35,800 --> 00:03:39,233
But here I'm going to put in command
by pressing command plus shift plus C.

72
00:03:39,433 --> 00:03:40,500
Here we go.

73
00:03:40,500 --> 00:03:43,933
And now when we have to do is to add this
you know library

74
00:03:44,500 --> 00:03:48,066
random forest
to actually automatically select

75
00:03:48,066 --> 00:03:51,066
the box here to import
automatically the random forest package

76
00:03:51,200 --> 00:03:54,166
when we execute the whole code
or the section.

77
00:03:54,166 --> 00:03:55,566
So that's important.

78
00:03:55,566 --> 00:03:58,066
And now time to build the regressor.

79
00:03:58,066 --> 00:03:59,200
So let's do it.

80
00:03:59,200 --> 00:04:04,300
We're going to call regressor regressor
as usual to keep things simple and equals.

81
00:04:04,633 --> 00:04:07,500
And now the function that we're going
to use is also

82
00:04:07,500 --> 00:04:10,733
random forest written the same.

83
00:04:10,966 --> 00:04:13,166
And so now let's add some parenthesis.

84
00:04:13,166 --> 00:04:17,433
And now let's press
F1 to have a look at the arguments okay.

85
00:04:17,433 --> 00:04:18,866
So the arguments are here.

86
00:04:18,866 --> 00:04:21,400
And the first argument is data.

87
00:04:21,400 --> 00:04:24,400
But as you can see it specifies
that it's an optional dataframe.

88
00:04:24,400 --> 00:04:27,566
And we could use this argument
to build our regressor.

89
00:04:27,566 --> 00:04:31,666
But we're going to use the main arguments
to specify, you know, the independent

90
00:04:31,666 --> 00:04:34,900
variables on one side and the dependent
variable and another side.

91
00:04:35,266 --> 00:04:39,233
And to do this we're going to use these
two arguments x and y.

92
00:04:39,333 --> 00:04:44,300
So x will contain the matrix and features
that is the independent variables.

93
00:04:44,533 --> 00:04:48,700
And y will contain the dependent
variable vector that is the salary column.

94
00:04:49,333 --> 00:04:51,466
So let's first input these two arguments.

95
00:04:51,466 --> 00:04:54,300
So the first argument is x equals.

96
00:04:54,300 --> 00:04:57,466
And so we have several ways
to take our independent variables.

97
00:04:57,766 --> 00:05:01,000
So one of the way is to take our data
set here.

98
00:05:01,500 --> 00:05:05,500
And then choose the right columns
of the independent variables.

99
00:05:05,833 --> 00:05:08,700
And you know our data set is composed
of two columns.

100
00:05:08,700 --> 00:05:12,000
The first column indexed by one
which is the independent variable column.

101
00:05:12,300 --> 00:05:16,600
And the second column indexed by two
which is our dependent variable column.

102
00:05:17,066 --> 00:05:20,700
So here we need index one because
we want to take the independent variable.

103
00:05:21,566 --> 00:05:22,433
All right.

104
00:05:22,433 --> 00:05:23,433
Now next argument.

105
00:05:23,433 --> 00:05:26,433
The next argument is y
the dependent variable vector.

106
00:05:26,466 --> 00:05:30,900
And now as you can see
y is expected to be a response vector.

107
00:05:30,900 --> 00:05:32,266
It's actually a vector.

108
00:05:32,266 --> 00:05:34,800
And here it expected to have a data frame.

109
00:05:34,800 --> 00:05:39,666
So by using this one index into brackets
here I actually import a data frame.

110
00:05:39,666 --> 00:05:43,333
But here to get a vector
I actually need to use another trick,

111
00:05:43,333 --> 00:05:46,200
another technique
which is in know to use a dollar sign.

112
00:05:46,200 --> 00:05:49,400
And then the name of the column,
which is of course salary.

113
00:05:50,366 --> 00:05:53,366
And that will give me a vector.
