1
00:00:00,370 --> 00:00:01,530
Now, Amazon Comprehend

2
00:00:01,530 --> 00:00:03,820
is a very simple service from an exam perspective.

3
00:00:03,820 --> 00:00:05,610
All you need to know is that Comprehend is

4
00:00:05,610 --> 00:00:06,540
to comprehend stuff,

5
00:00:06,540 --> 00:00:09,928
so it is for Natural Language Processing, or NLP.

6
00:00:09,928 --> 00:00:11,870
So anytime you will see NLP on the exam,

7
00:00:11,870 --> 00:00:14,020
you would think Amazon Comprehend.

8
00:00:14,020 --> 00:00:17,830
Now what it is, it is a fully managed and serverless service

9
00:00:17,830 --> 00:00:20,040
and it will use machine learning to find insights

10
00:00:20,040 --> 00:00:22,840
and relationships in your text.

11
00:00:22,840 --> 00:00:24,758
For example, it can understand what is the language

12
00:00:24,758 --> 00:00:27,490
of the text, it can extract key phrases,

13
00:00:27,490 --> 00:00:30,460
places, people, brands, or events from that text.

14
00:00:30,460 --> 00:00:33,340
It can understand, do a sentiment analysis to understand

15
00:00:33,340 --> 00:00:37,570
how positive or negative the text you're analyzing is.

16
00:00:37,570 --> 00:00:40,320
It can also analyze a text using tokenization

17
00:00:40,320 --> 00:00:41,290
and parts of speech.

18
00:00:41,290 --> 00:00:42,900
It also has audio.

19
00:00:42,900 --> 00:00:45,190
And finally, it can organize a collection

20
00:00:45,190 --> 00:00:48,720
of text files by topic, and find topics itself.

21
00:00:48,720 --> 00:00:51,850
So, Comprehend is really about getting a lot of data in

22
00:00:51,850 --> 00:00:53,010
and then Comprehend will do the rest

23
00:00:53,010 --> 00:00:55,660
to try to understand the meaning of that data.

24
00:00:55,660 --> 00:00:58,067
So it's about taking text or unstructured data

25
00:00:58,067 --> 00:01:00,900
and structuring it around these features.

26
00:01:00,900 --> 00:01:03,500
So some sample use cases of Natural Language Processing,

27
00:01:03,500 --> 00:01:06,770
or NLP, are for example, to analyze customer interaction.

28
00:01:06,770 --> 00:01:09,640
So say you have a bunch of customers sending you emails

29
00:01:09,640 --> 00:01:11,800
and you want to understand overall based

30
00:01:11,800 --> 00:01:14,190
on your support service what leads to a positive

31
00:01:14,190 --> 00:01:16,260
or negative experience from your customers.

32
00:01:16,260 --> 00:01:18,840
So we use Comprehend to extract these features

33
00:01:18,840 --> 00:01:20,630
and then you will have the business insight

34
00:01:20,630 --> 00:01:21,694
and then you can improve your business,

35
00:01:21,694 --> 00:01:24,050
thanks to that analysis.

36
00:01:24,050 --> 00:01:26,020
You can also, for example, create

37
00:01:26,020 --> 00:01:29,080
and group articles by topics that Comprehend will uncover.

38
00:01:29,080 --> 00:01:30,520
So imagine you have a lot of articles

39
00:01:30,520 --> 00:01:32,450
and you want to group them together instead

40
00:01:32,450 --> 00:01:34,740
of going to them one by one, you would feed them

41
00:01:34,740 --> 00:01:36,920
into Comprehend and then it would output you

42
00:01:36,920 --> 00:01:38,810
the topics you need to group them by.

43
00:01:38,810 --> 00:01:39,643
So that's it.

44
00:01:39,643 --> 00:01:41,500
So again, a bit more information than what you need,

45
00:01:41,500 --> 00:01:44,060
but Comprehend from an exam perspective is

46
00:01:44,060 --> 00:01:45,760
for Natural Language Processing.

47
00:01:45,760 --> 00:01:46,593
So hope you liked it,

48
00:01:46,593 --> 00:01:48,520
and I will see you in the next lecture.