﻿1
00:00:00,133 --> 00:00:02,400
Hello and welcome to this art tutorial.

2
00:00:02,400 --> 00:00:04,600
So we already built two regression models.

3
00:00:04,600 --> 00:00:08,033
It was a simple linear regression
and the multiple linear regression.

4
00:00:08,433 --> 00:00:12,300
And today in this tutorial
we will start with polynomial regression.

5
00:00:13,066 --> 00:00:16,233
So as usual let's prepare the workspace
by setting the right folder

6
00:00:16,233 --> 00:00:17,633
as working directory.

7
00:00:17,633 --> 00:00:19,966
So right now I'm going to my Machine
learning A-Z folder.

8
00:00:19,966 --> 00:00:23,366
Part two regression
then section polynomial regression.

9
00:00:23,600 --> 00:00:24,966
And here is the right folder.

10
00:00:24,966 --> 00:00:27,733
Make sure that you have the position
salary csv file.

11
00:00:27,733 --> 00:00:30,066
And you can click on this
more button here.

12
00:00:30,066 --> 00:00:32,366
And then set as working directory.

13
00:00:32,366 --> 00:00:36,800
And now let's start with the usual first
step of making a machine learning model.

14
00:00:37,333 --> 00:00:40,633
So this first step is
of course the data pre-processing step.

15
00:00:40,900 --> 00:00:42,266
So to be efficient I'm

16
00:00:42,266 --> 00:00:45,900
going to go to my data pre-processing
template that we made in part one.

17
00:00:46,266 --> 00:00:48,366
I'm going to select
everything in the template

18
00:00:49,433 --> 00:00:51,033
copy and

19
00:00:51,033 --> 00:00:54,033
paste it in my polynomial regression
model.

20
00:00:55,000 --> 00:00:56,866
This way okay.

21
00:00:56,866 --> 00:00:58,833
And now
we just need to change a few things.

22
00:00:58,833 --> 00:01:01,866
So the first thing we need to change
is the name of the data set.

23
00:01:02,266 --> 00:01:03,466
So it's not data here.

24
00:01:03,466 --> 00:01:06,466
The data was the name of the data
set in the pre-processing part.

25
00:01:06,466 --> 00:01:08,066
But now in this regression part

26
00:01:08,066 --> 00:01:11,100
the name of the data set is position
underscore salaries.

27
00:01:11,400 --> 00:01:14,400
So here we'll write position

28
00:01:14,666 --> 00:01:17,000
underscore salaries.

29
00:01:17,000 --> 00:01:19,200
And now let's select this line

30
00:01:19,200 --> 00:01:22,200
and execute
to have a look at our data set.

31
00:01:22,800 --> 00:01:23,166
Okay.

32
00:01:23,166 --> 00:01:26,166
And now let's see
what will be our machine learning mission.

33
00:01:26,800 --> 00:01:30,100
So we are a human resource team
working for a big company.

34
00:01:30,400 --> 00:01:33,400
And we are about to hire a new employee
in this company.

35
00:01:33,433 --> 00:01:36,466
So this new employee seems to be great
a good fit for the job.

36
00:01:36,633 --> 00:01:40,333
And we are about to make an offer
to this potential new employee.

37
00:01:40,933 --> 00:01:42,566
And now it's time to negotiate.

38
00:01:42,566 --> 00:01:43,500
Negotiate on

39
00:01:43,500 --> 00:01:47,400
what is going to be the future salary
of this new employee in the company.

40
00:01:48,100 --> 00:01:50,033
And so,
at the beginning of the negotiation,

41
00:01:50,033 --> 00:01:54,033
this new employee is telling
that he's had 20 plus years of experience

42
00:01:54,333 --> 00:01:58,400
and eventually earns 160 K
annual salary in its previous company.

43
00:01:58,766 --> 00:02:02,433
So this employee is asking for
at least more than 160 K.

44
00:02:03,066 --> 00:02:06,166
However, there is someone in the HR team
that is kind of a control

45
00:02:06,166 --> 00:02:09,166
freak and always fantasized
about being a detective.

46
00:02:09,333 --> 00:02:13,000
So suddenly decides to call the previous
employee to check that info.

47
00:02:13,033 --> 00:02:16,000
You know that info about the previous
160 K

48
00:02:16,000 --> 00:02:19,000
annual salary of this future
potential new employee.

49
00:02:19,366 --> 00:02:22,300
But unfortunately,
all the info that this person

50
00:02:22,300 --> 00:02:25,300
manages to get are these info here.

51
00:02:25,366 --> 00:02:28,833
The simple table of salaries
for these ten different positions

52
00:02:29,066 --> 00:02:30,600
in the previous company.

53
00:02:30,600 --> 00:02:35,000
So this HR member of the team runs
a simple analysis on Excel or Google

54
00:02:35,000 --> 00:02:38,833
Sheets, and actually observes
that there is a nonlinear relationship

55
00:02:39,000 --> 00:02:42,900
between these position levels
and their associated salaries.

56
00:02:43,866 --> 00:02:45,533
However, and moreover,

57
00:02:45,533 --> 00:02:48,666
this HR guy could get another
very relevant info.

58
00:02:49,300 --> 00:02:53,433
This other relevant info is that this
new employee that is about to be hired

59
00:02:53,700 --> 00:02:57,366
has been a region manager for two years
now in the previous company.

60
00:02:57,866 --> 00:03:00,600
And that usually takes an average
four years to jump

61
00:03:00,600 --> 00:03:03,600
from being a region manager to a partner.

62
00:03:03,600 --> 00:03:06,833
So this new employee about to be hired

63
00:03:06,833 --> 00:03:10,133
was kind of halfway between level
six and level seven.

64
00:03:10,133 --> 00:03:13,133
And therefore we can say he was level 6.5.

65
00:03:13,500 --> 00:03:14,566
So now this H.R.

66
00:03:14,566 --> 00:03:17,566
Guy is getting all excited
because he's telling to the team

67
00:03:17,566 --> 00:03:20,700
that he can build a bluffing detector
using regression models

68
00:03:20,933 --> 00:03:23,933
to predict if this new employee
is bluffing about its salary.

69
00:03:24,400 --> 00:03:27,066
So at the beginning,
the team finds it a little weird,

70
00:03:27,066 --> 00:03:30,066
but it's kind of curious
to see what's going to happen.

71
00:03:30,133 --> 00:03:31,866
So here is the mission.

72
00:03:31,866 --> 00:03:35,133
This new employee that is about to be
hired is telling

73
00:03:35,133 --> 00:03:38,833
that its annual salary was 160
K in the previous company.

74
00:03:39,366 --> 00:03:43,433
Let's use polynomial regression
to build a bluffing detector to predict

75
00:03:43,433 --> 00:03:44,766
if it's truth or bluff.
