0
1
00:00:00,240 --> 00:00:03,100
Next topic would be unsupervised learning.
1

2
00:00:03,540 --> 00:00:08,670
This is where those labels are omitted, rather, unlabeled
2

3
00:00:08,760 --> 00:00:14,740
It is form of machine learning rather than being trained with sample data,
3

4
00:00:15,000 --> 00:00:23,550
the machine learning system finds structures and patterns in the data on its own. With unsupervised
4

5
00:00:23,550 --> 00:00:24,480
algorithms,
5

6
00:00:24,490 --> 00:00:29,220
you still don't know what you want to get out of the model yet!
6

7
00:00:29,550 --> 00:00:36,630
You have a bunch of data, but you really don't know how to categorize them, how to put them in different
7

8
00:00:36,630 --> 00:00:40,510
groups and you want the machine to do it for you in this case.
8

9
00:00:40,510 --> 00:00:48,330
So you normalize your data into a format that makes sense to compare them and then let the model work
9

10
00:00:48,570 --> 00:00:52,500
its magic and try to find some of these relationships.
10

11
00:00:52,680 --> 00:00:59,190
One of the special characteristic of this model is that while the model can suggest different ways to
11

12
00:00:59,190 --> 00:01:07,230
categorize or order your data, it's also at the end up to you to make the furthermore research on those
12

13
00:01:07,230 --> 00:01:09,240
datasets and make a decision.
13

14
00:01:09,660 --> 00:01:18,000
For example, after processing all the data related to all the product's user with an unsupervised algorithm,
14

15
00:01:18,240 --> 00:01:22,980
it might come up with a way to group your users into two groups.
15

16
00:01:23,490 --> 00:01:32,310
After inspecting and comparing these two groups, you might realize that Group A is in a geographic
16

17
00:01:32,310 --> 00:01:39,030
location of Asia, for example, and Group B is in a geographic location of Europe.
17

18
00:01:39,030 --> 00:01:43,100
Later, it's up to you to make furthermore decisions on this system.
18

19
00:01:43,560 --> 00:01:50,010
So whenever we have some kind of big data and we don't know how to group there, we just ask the machine
19

20
00:01:50,190 --> 00:01:55,590
with unsupervised learning to come up with somehow grouping the data.