1
00:00:00,260 --> 00:00:01,220
Hey, welcome back.

2
00:00:01,220 --> 00:00:08,780
In this video, I'll introduce you to the three types of recommendation systems you are using as a user

3
00:00:08,780 --> 00:00:12,110
every day when you're browsing Netflix or Amazon.

4
00:00:12,110 --> 00:00:15,690
And then we're going to build these three types in the next videos.

5
00:00:15,710 --> 00:00:23,360
The first type of a recommendation system is known as popularity based recommendation system, which

6
00:00:23,360 --> 00:00:32,090
means that if you are a user and you are browsing Netflix but you are not logged in, so Netflix doesn't

7
00:00:32,090 --> 00:00:36,890
know anything about you yet, so it gives you a front page.

8
00:00:36,890 --> 00:00:46,940
So I'm in Netflix.com browse genre, so I'm just browsing movies here so Netflix doesn't know anything

9
00:00:46,940 --> 00:00:52,490
about you, so it will give you a list of the most popular movies.

10
00:00:52,850 --> 00:00:56,630
So it is recommending you based on popularity.

11
00:00:56,630 --> 00:01:00,630
What is being liked from the majority of people?

12
00:01:00,810 --> 00:01:02,310
Same for Amazon.

13
00:01:02,310 --> 00:01:04,980
So it gives you on the front page.

14
00:01:04,980 --> 00:01:13,500
If you are not logged in, it gives you the most popular products, which means that since Amazon doesn't

15
00:01:13,500 --> 00:01:20,280
know anything about you by showing you the most popular products, it's more likely that you will buy

16
00:01:20,280 --> 00:01:21,480
something.

17
00:01:22,020 --> 00:01:24,630
This is the first type of recommendation system.

18
00:01:24,630 --> 00:01:32,460
But then when if you log in to Amazon or to Netflix and your user account has some history, so perhaps

19
00:01:32,460 --> 00:01:39,030
you watched some movies already, so you have some minutes watched in the Netflix database or you purchase

20
00:01:39,030 --> 00:01:41,430
something on Amazon already.

21
00:01:41,460 --> 00:01:48,810
So Amazon and Netflix, they have a database about you and they know what you watched and what you rated

22
00:01:48,810 --> 00:01:50,520
as high and low.

23
00:01:51,390 --> 00:01:55,380
Then Amazon also knows about other users, right?

24
00:01:55,380 --> 00:01:57,270
So it has a database of users.

25
00:01:57,270 --> 00:02:05,250
So by knowing what you like and by knowing what the other user likes, then it can show you potential

26
00:02:05,250 --> 00:02:13,620
products or movies you may want to consume because that other user who liked and watched the same thing,

27
00:02:13,800 --> 00:02:18,900
they enjoyed these potential next movies and products.

28
00:02:18,900 --> 00:02:22,980
So this is known as collaborative filtering.

29
00:02:22,980 --> 00:02:26,910
So this was the second type of a recommendation system.

30
00:02:26,910 --> 00:02:32,280
So so far we have popularity based and collaborative filtering.

31
00:02:32,310 --> 00:02:40,320
The third type of recommendation system is known as content based filtering, which means that if we,

32
00:02:40,320 --> 00:02:47,220
for example, go inside one of these movies, let's let's click on this, this here.

33
00:02:48,850 --> 00:02:54,370
Now, Netflix will recommend similar movies down here below, right?

34
00:02:54,370 --> 00:03:02,560
Because it knows what this movie is about since it has a database of the movie title, the movie description

35
00:03:02,560 --> 00:03:06,670
and other information about that movie, the genre, etcetera.

36
00:03:06,670 --> 00:03:09,580
And then it has the same data for other movies.

37
00:03:09,580 --> 00:03:18,670
So it compares these data and it finds a similarity, it calculates a similarity coefficient, and then

38
00:03:18,670 --> 00:03:24,220
it shows you the movies which are more most similar to that movie.

39
00:03:24,310 --> 00:03:30,220
And this recommendation system is known as content based filtering.

40
00:03:30,220 --> 00:03:37,570
So these are the three types of recommendation systems, popularity based, collaborative and content

41
00:03:37,570 --> 00:03:38,680
based filtering.

42
00:03:38,680 --> 00:03:42,910
Now in the next videos, we're going to implement all these three.

43
00:03:42,910 --> 00:03:46,300
We're going to start with a movie database.

44
00:03:46,300 --> 00:03:50,800
I'm going to introduce you to this database, to this data in the next video.

45
00:03:51,250 --> 00:03:52,900
So that's the first step.

46
00:03:52,900 --> 00:03:57,550
When you deal with data science and machine learning, you want to get to know your data.

47
00:03:57,970 --> 00:04:02,380
And so I'll give you the data and we'll look through them together in the next video.

48
00:04:02,440 --> 00:04:03,760
See you in the next video.