1
00:00:00,200 --> 00:00:02,466
Hello and welcome to this art tutorial.

2
00:00:02,466 --> 00:00:05,166
So now let's proceed to the next step
of the cleaning process,

3
00:00:05,166 --> 00:00:08,166
which is about removing
all the numbers from the reviews.

4
00:00:08,433 --> 00:00:12,066
So we will just copy this line
and paste it below,

5
00:00:12,333 --> 00:00:15,900
because you'll see that now during
the next steps of the cleaning process

6
00:00:15,900 --> 00:00:17,066
is going to be very easy

7
00:00:17,066 --> 00:00:21,366
because we will always use the same line
here to update the corpus.

8
00:00:21,700 --> 00:00:24,700
And at each new step,
we are going to apply the right

9
00:00:24,700 --> 00:00:28,600
transformation that we want to do for the
particular step of the cleaning process

10
00:00:28,866 --> 00:00:31,500
to the whole corpus,
thanks to the tmp function.

11
00:00:31,500 --> 00:00:35,400
So basically what we just need to do
is replace this content transformer

12
00:00:35,400 --> 00:00:39,766
with two layers input by remove numbers.

13
00:00:40,066 --> 00:00:41,666
And that's how we apply this

14
00:00:41,666 --> 00:00:45,300
remove numbers function to the reviews
in the corpus through the tmp function.

15
00:00:45,533 --> 00:00:50,166
And that will remove all the numbers
of all the 1000 reviews in the corpus.

16
00:00:50,466 --> 00:00:51,700
So let's check it out.

17
00:00:51,700 --> 00:00:52,433
To check that out.

18
00:00:52,433 --> 00:00:55,600
We cannot do it with this first review
because this first review doesn't

19
00:00:55,600 --> 00:00:58,666
contain any number,
so nothing will be removed here.

20
00:00:58,666 --> 00:01:00,733
But I had a look at the data set,

21
00:01:00,733 --> 00:01:04,766
and there is the review 841
that contains a number.

22
00:01:05,066 --> 00:01:07,100
Let's have a look at this review.

23
00:01:07,100 --> 00:01:10,533
To do this we will use the same line
as we used to look at the first review.

24
00:01:10,833 --> 00:01:14,633
So I'm pressing the up arrow
here to take this line of code

25
00:01:14,700 --> 00:01:16,866
that gives us access to the reviews.

26
00:01:16,866 --> 00:01:21,966
And inputting the index
one here will input the 841 index.

27
00:01:22,400 --> 00:01:25,100
And now let's press enter
to have a look at the review.

28
00:01:25,100 --> 00:01:27,866
And the review is for 40 bucks ahead.

29
00:01:27,866 --> 00:01:29,700
I really expect better food.

30
00:01:29,700 --> 00:01:30,933
So that's a negative review.

31
00:01:30,933 --> 00:01:34,800
And what should be highlighted
in this review is this 40 number here

32
00:01:34,833 --> 00:01:37,900
which we want to see
if it's going to be removed.

33
00:01:38,100 --> 00:01:42,133
Once we apply this remove numbers function
to the reviews of the corpus,

34
00:01:42,333 --> 00:01:46,500
so that all the numbers in the reviews
can be removed thanks to the tmp function.

35
00:01:46,766 --> 00:01:48,200
So let's check it out.

36
00:01:48,200 --> 00:01:51,200
We will select this line and execute.

37
00:01:51,666 --> 00:01:52,466
Done.

38
00:01:52,466 --> 00:01:56,100
And now let's have a look at this 841
review.

39
00:01:56,500 --> 00:01:59,466
So I'm pressing the up arrow twice

40
00:01:59,466 --> 00:02:03,233
to get back to this line of code
giving us the written review.

41
00:02:03,433 --> 00:02:05,333
So now the corpus is updated.

42
00:02:05,333 --> 00:02:08,233
Let's see if the number 40 disappeared

43
00:02:08,233 --> 00:02:11,100
and it did four bucks ahead.

44
00:02:11,100 --> 00:02:12,933
I really expect better food.

45
00:02:12,933 --> 00:02:14,033
40 disappeared.

46
00:02:14,033 --> 00:02:16,566
And that's exactly what we wanted.
So great.

47
00:02:16,566 --> 00:02:21,166
Basically all the numbers are now removed
from the reviews in the corpus.

48
00:02:21,433 --> 00:02:24,733
Next step done, and we are ready
to move on to the next step,

49
00:02:25,033 --> 00:02:29,666
which will be about removing
any kind of punctuation in the reviews,

50
00:02:29,966 --> 00:02:33,733
because of course, in our sparse matrix,
in the end, we don't want to get a column

51
00:02:33,733 --> 00:02:37,700
for a comma or another column
for a colon, or another column

52
00:02:37,700 --> 00:02:41,700
for a dot, or for a semicolon,
or any kind of punctuation.

53
00:02:42,033 --> 00:02:45,333
Of course, we only want
to create some columns for relevant words.

54
00:02:45,333 --> 00:02:48,200
That will help the machine
learning classification algorithm

55
00:02:48,200 --> 00:02:49,300
to see the correlations

56
00:02:49,300 --> 00:02:51,533
between the presence of the words
and the outcome,

57
00:02:51,533 --> 00:02:54,000
whether the review is positive
or negative.

58
00:02:54,000 --> 00:02:54,333
All right.

59
00:02:54,333 --> 00:02:56,100
So let's do that in the next tutorial.

60
00:02:56,100 --> 00:02:57,733
And until then enjoy machine learning.