1
00:00:00,533 --> 00:00:02,633
Hello and welcome to the Machine
Learning A to Z course.

2
00:00:02,633 --> 00:00:05,300
Super excited
to have you back here on board.

3
00:00:05,300 --> 00:00:07,233
You might be wondering
what's up with the timer?

4
00:00:07,233 --> 00:00:11,466
Well, this tutorial is all about
getting excited about machine learning,

5
00:00:11,766 --> 00:00:14,500
and I've set the challenge of showing you

6
00:00:14,500 --> 00:00:17,600
the power of machine
learning in under five minutes.

7
00:00:18,000 --> 00:00:19,966
So let's dive straight into it.

8
00:00:19,966 --> 00:00:22,933
You are a data scientist
working for a car company,

9
00:00:22,933 --> 00:00:24,333
and you've been given this dataset

10
00:00:24,333 --> 00:00:27,966
with ages and estimated salaries
of potential customers.

11
00:00:28,300 --> 00:00:33,600
Your task is to predict which ones of
these customers are more likely

12
00:00:33,733 --> 00:00:39,633
to purchase a car, based on a campaign
that the sales division will be running.

13
00:00:40,033 --> 00:00:43,566
The good news is that the sales division
also gave you this data set,

14
00:00:43,766 --> 00:00:47,833
which is data from a previous campaign,
a very similar campaign

15
00:00:47,833 --> 00:00:51,900
that ran in the past, which also has ages
and estimated salaries of customers.

16
00:00:52,166 --> 00:00:56,700
But it has an additional column which says
whether that customer purchased the car

17
00:00:57,000 --> 00:01:00,566
that they were advertised
or whether they didn't purchase a car.

18
00:01:00,866 --> 00:01:04,500
So this data set is the one
we're going to use to build a model,

19
00:01:04,500 --> 00:01:07,666
and then we're going to apply that model
to this data set.

20
00:01:07,666 --> 00:01:10,200
And we're going to be using
a logistic regression model.

21
00:01:10,200 --> 00:01:13,200
Excited.
So my let's dive straight into it.

22
00:01:13,500 --> 00:01:14,700
We're going to be using Python.

23
00:01:14,700 --> 00:01:18,233
And we're going to be working in
Google Colab which is a very popular tool.

24
00:01:18,600 --> 00:01:20,866
As you can see we have some code
pre-written here.

25
00:01:20,866 --> 00:01:24,533
And we're going to walk through this code
and run it step by step.

26
00:01:24,533 --> 00:01:27,533
And you'll see the model getting built,
which is very exciting.

27
00:01:27,700 --> 00:01:30,700
Now let's take the first step
and we'll load the data sets.

28
00:01:30,700 --> 00:01:31,533
Here are the data sets.

29
00:01:31,533 --> 00:01:32,566
We're going to click over here.

30
00:01:32,566 --> 00:01:34,700
And I'm going to load them up.

31
00:01:34,700 --> 00:01:39,033
All right so the data sets are loaded
and we can proceed to our code. Now.

32
00:01:40,366 --> 00:01:41,266
So there we go.

33
00:01:41,266 --> 00:01:44,166
First thing we're going to do
is we're going to import the libraries.

34
00:01:44,166 --> 00:01:47,833
So we're going to open this up
run these libraries.

35
00:01:47,833 --> 00:01:50,833
And this will allow us to work
with the data.

36
00:01:51,000 --> 00:01:54,800
Next we're going to import the data
set into our code so we can work with it

37
00:01:54,800 --> 00:01:55,066
there.

38
00:01:55,066 --> 00:01:56,100
There's you can see there's

39
00:01:56,100 --> 00:01:59,400
the age estimated salary
and whether or not a person purchased.

40
00:01:59,400 --> 00:02:01,366
So that's a previous campaign.

41
00:02:01,366 --> 00:02:04,533
Then we're going to plot the data set
and see what it's all about.

42
00:02:04,900 --> 00:02:07,966
Here we can see the blue dots
represent people who purchase.

43
00:02:07,966 --> 00:02:12,600
They're usually at a higher age
or at a higher salary.

44
00:02:12,933 --> 00:02:15,400
And people who didn't
purchase are the red dots.

45
00:02:15,400 --> 00:02:17,566
And of course
there's a mix over there as well.

46
00:02:17,566 --> 00:02:19,066
Now we're going to apply feature scaling.

47
00:02:19,066 --> 00:02:23,000
We won't go into detail on what this is
all about is just a very important step,

48
00:02:23,000 --> 00:02:26,166
and you'll learn more about it
inside the course for sure.

49
00:02:27,033 --> 00:02:30,000
Next, we're going to train
the logistic regression model.

50
00:02:30,000 --> 00:02:31,033
And there we go.

51
00:02:31,033 --> 00:02:33,133
logistic regression model is trained.

52
00:02:33,133 --> 00:02:36,100
And logistic regression is just
one of the models, a very popular one.

53
00:02:36,100 --> 00:02:39,233
But definitely they are the ones which
again, you will learn inside the course.

54
00:02:39,766 --> 00:02:42,733
We're going to visual the visualize
the model results.

55
00:02:42,733 --> 00:02:45,766
And this will allow us to see
what is going on

56
00:02:45,766 --> 00:02:48,766
in our data
and how the model is being applied to it.

57
00:02:48,900 --> 00:02:49,800
So there we go.

58
00:02:49,800 --> 00:02:53,400
There's our logistic regression
model applied to the data.

59
00:02:53,400 --> 00:02:56,766
As you can see, it's
saying that anything above this blue line

60
00:02:56,800 --> 00:02:59,700
means that the person will purchase
or has purchased a car.

61
00:02:59,700 --> 00:03:02,700
Anything below
the line means that they won't.

62
00:03:02,800 --> 00:03:06,300
and of course there are some mismatches
with the blue and red dots.

63
00:03:06,300 --> 00:03:08,800
And that's okay,
because no model is perfect.

64
00:03:08,800 --> 00:03:12,433
Those are just errors
which are totally normal for models.

65
00:03:12,433 --> 00:03:15,500
And in this course you will learn
how to pick the right models

66
00:03:15,500 --> 00:03:19,200
for the right applications
and to minimize those errors.

67
00:03:19,600 --> 00:03:22,600
Next we're going to import
the new project data.

68
00:03:22,733 --> 00:03:24,166
And there it is.

69
00:03:24,166 --> 00:03:27,033
That's our new data
that we need to predict.

70
00:03:27,033 --> 00:03:30,266
we're going to apply our model
to make those predictions.

71
00:03:30,533 --> 00:03:36,400
And we will then visualize the predictions
to see, the results.

72
00:03:36,400 --> 00:03:37,133
So there we go.

73
00:03:37,133 --> 00:03:39,200
Those are our results over here.

74
00:03:39,200 --> 00:03:44,566
And as we can see the model is saying
that anything above this blue line,

75
00:03:44,900 --> 00:03:47,566
that those people will are likely

76
00:03:47,566 --> 00:03:50,733
to purchase the car below
the blue line are not likely to purchase.

77
00:03:51,000 --> 00:03:56,100
So our marketing department can just
simply target the people in the blue area.

78
00:03:56,366 --> 00:04:00,533
And therefore save on costs
and optimize their efforts,

79
00:04:00,533 --> 00:04:03,533
focus their efforts,
and get the best return on investment.

80
00:04:03,633 --> 00:04:04,200
So there we go.

81
00:04:04,200 --> 00:04:07,033
That's how easy it is to apply machine
learning models.

82
00:04:07,033 --> 00:04:09,500
As you can see,
this took us under five minutes.

83
00:04:09,500 --> 00:04:12,000
And of course
this was a simplified approach.

84
00:04:12,000 --> 00:04:13,200
We didn't do many of the steps,

85
00:04:13,200 --> 00:04:16,433
like splitting the data
into a training set and a test set.

86
00:04:16,566 --> 00:04:20,900
We didn't, build a confusion
matrix or calculate accuracy ratios

87
00:04:21,266 --> 00:04:25,466
and lots of other nuances, all of which
you will learn inside the course.

88
00:04:25,633 --> 00:04:28,600
But in a nutshell,
this is how machine learning works.

89
00:04:28,600 --> 00:04:32,100
And this is the power of machine learning,

90
00:04:32,100 --> 00:04:35,433
which already soon
you will be able to apply in your career.

91
00:04:35,866 --> 00:04:38,833
I hope you're excited
and I can't wait to see you inside

92
00:04:38,833 --> 00:04:42,533
the course, where we will learn a lot
and have lots of fun along the way.

93
00:04:42,866 --> 00:04:44,900
And until next
time, enjoy machine learning!