1
00:00:00,300 --> 00:00:01,066
Perfect.

2
00:00:01,066 --> 00:00:02,566
And now, final step.

3
00:00:02,566 --> 00:00:04,633
You know what you have to do, right?

4
00:00:04,633 --> 00:00:07,700
We have to take first
our classifier and then.

5
00:00:07,700 --> 00:00:12,066
Well, now what we want to do is to train
our classifier on the training set.

6
00:00:12,300 --> 00:00:15,966
Because remember that
this line of code only builds

7
00:00:16,166 --> 00:00:19,166
the logistic regression model
but doesn't train it yet.

8
00:00:19,200 --> 00:00:22,333
And this next line of code is indeed
the final step

9
00:00:22,333 --> 00:00:25,633
where we train our classifier
on the training set.

10
00:00:25,966 --> 00:00:30,533
And remember to do this,
we have to call the fit method,

11
00:00:30,733 --> 00:00:34,566
which takes as input two entities,
you know, two sets of data.

12
00:00:34,566 --> 00:00:37,933
The first one is the matrix of features
of the training set,

13
00:00:38,266 --> 00:00:41,700
and the second one is the dependent
variable vector of the training set, where

14
00:00:41,700 --> 00:00:45,300
of course, all the purchase decisions
of the dependent variable vector

15
00:00:45,633 --> 00:00:49,800
correspond to the same customers
of the matrix of features.

16
00:00:49,900 --> 00:00:50,700
All right.

17
00:00:50,700 --> 00:00:53,800
So what we had to input here was simply X

18
00:00:54,400 --> 00:00:57,333
train for the matrix of features
of the training set.

19
00:00:57,333 --> 00:01:00,333
And then y train for

20
00:01:00,333 --> 00:01:03,333
the dependent variable vector
of the same training set.

21
00:01:03,833 --> 00:01:06,266
And there you go
my friends. Congratulations.

22
00:01:06,266 --> 00:01:10,200
If you obtained this
and also congratulations if you tried.

23
00:01:10,500 --> 00:01:14,266
Because indeed that's how you build
the logistic regression model.

24
00:01:14,266 --> 00:01:15,000
So there you go.

25
00:01:15,000 --> 00:01:18,000
That's one extra machinery model
in your toolkit.

26
00:01:18,166 --> 00:01:20,333
And you're going to have many more.

27
00:01:20,333 --> 00:01:23,800
And don't worry
I will train you on how to select

28
00:01:23,800 --> 00:01:26,800
the best one for any data set.

29
00:01:27,066 --> 00:01:29,500
Okay. Great. So let's run this cell.

30
00:01:29,500 --> 00:01:32,766
This will indeed build
and train the logistic regression

31
00:01:32,766 --> 00:01:36,600
model on your training set composed of X
and Y train.

32
00:01:36,866 --> 00:01:39,100
And here, by the way,
you see all the parameters

33
00:01:39,100 --> 00:01:41,800
of the logistic regression model
which you can tune.

34
00:01:41,800 --> 00:01:45,366
The most famous one is C
which is inverse of the regularization

35
00:01:45,366 --> 00:01:50,133
strength, meaning that the smaller is c,
the stronger will be the regularization,

36
00:01:50,133 --> 00:01:53,700
and therefore the more
it will protect you from overfitting.

37
00:01:54,100 --> 00:01:56,833
Right here
we'll just keep the default value of one.

38
00:01:56,833 --> 00:01:57,700
And there you go.

39
00:01:57,700 --> 00:02:01,433
We're ready to move on to the next step
predicting a new result.

40
00:02:01,633 --> 00:02:06,600
And so now new exercise for you
because we already did it with regression

41
00:02:06,600 --> 00:02:10,700
I taught you before how to, you know,
take your regressor and then call

42
00:02:10,700 --> 00:02:14,133
the predict method to predict the result
of a single observation.

43
00:02:14,500 --> 00:02:17,633
And so here
the exercise that I would like you to do

44
00:02:18,066 --> 00:02:22,000
is to predict
the purchased decision of a single result,

45
00:02:22,200 --> 00:02:25,033
you know, the purchase decision
of a single customer.

46
00:02:25,033 --> 00:02:27,766
And I will tell you which one
I would like you to predict

47
00:02:27,766 --> 00:02:32,600
the purchased decision
of the first customer in the test set.

48
00:02:32,900 --> 00:02:36,900
Who you will see is 30 years old and earns

49
00:02:36,900 --> 00:02:40,633
an estimated salary of $87,000.

50
00:02:40,833 --> 00:02:42,900
So I would like you
to take these two inputs.

51
00:02:42,900 --> 00:02:46,400
You know, these two features
with these values, 30 years old

52
00:02:46,400 --> 00:02:49,500
and $87,000 as the estimated salary.

53
00:02:49,700 --> 00:02:55,066
And I would like you to predict whether
this customer has bought yes or no.

54
00:02:55,200 --> 00:02:58,566
The SUV, you have the answer and y test.

55
00:02:58,566 --> 00:03:01,233
You know,
you take the first result of Y test,

56
00:03:01,233 --> 00:03:05,166
and thus you'll be able to compare
your prediction to the real result

57
00:03:05,166 --> 00:03:08,166
to figure out
whether your prediction was correct.

58
00:03:08,466 --> 00:03:11,100
And so there you go.
Please do the exercise.

59
00:03:11,100 --> 00:03:14,400
Please try to do this on your own first
before we do it together

60
00:03:14,400 --> 00:03:17,800
in the next tutorial,
please predict the purchase decision

61
00:03:17,800 --> 00:03:21,900
of that first customer of the test set
and we will

62
00:03:22,200 --> 00:03:26,800
of course implement
the solution in the next tutorial.

63
00:03:27,433 --> 00:03:29,700
Until then, enjoy machine learning.