en_Intro to Recommender Systems.srt

0
00:00:00,750 --> 00:00:04,490
Hello, and welcome! In this video, we’ll be going through a

1
00:00:04,490 --> 00:00:09,800
quick introduction to recommendation systems. So, let’s get started.

2
00:00:09,800 --> 00:00:14,450
Even though peoples’ tastes may vary, they generally follow patterns.

3
00:00:14,450 --> 00:00:19,590
By that, I mean that there are similarities in the things that people tend to like … or

4
00:00:19,590 --> 00:00:24,590
another way to look at it, is that people tend to like things in the same category or

5
00:00:24,590 --> 00:00:30,280
things that share the same characteristics. For example, if you’ve recently purchased

6
00:00:30,280 --> 00:00:36,100
a book on Machine Learning in Python and you’ve enjoyed reading it, it’s very likely that

7
00:00:36,100 --> 00:00:40,800
you’ll also enjoy reading a book on Data Visualization.

8
00:00:40,800 --> 00:00:45,020
People also tend to have similar tastes to those of the people they’re close to in

9
00:00:45,020 --> 00:00:49,520
their lives. Recommender systems try to capture these patterns

10
00:00:49,520 --> 00:00:54,489
and similar behaviors, to help predict what else you might like.

11
00:00:54,489 --> 00:00:58,090
Recommender systems have many applications that I’m sure you’re already familiar

12
00:00:58,090 --> 00:01:01,120
with. Indeed, Recommender systems are usually at

13
00:01:01,120 --> 00:01:06,250
play on many websites. For example, suggesting books on Amazon and

14
00:01:06,250 --> 00:01:11,200
movies on Netflix. In fact, everything on Netflix’s website

15
00:01:11,200 --> 00:01:15,840
is driven by customer selection. If a certain movie gets viewed frequently

16
00:01:15,840 --> 00:01:21,770
enough, Netflix’s recommender system ensures that that movie gets an increasing number

17
00:01:21,770 --> 00:01:24,150
of recommendations.

18
00:01:24,150 --> 00:01:29,860
Another example can be found in a daily-use mobile app, where a recommender engine is

19
00:01:29,860 --> 00:01:35,300
used to recommend anything from where to eat, or, what job to apply to.

20
00:01:35,300 --> 00:01:42,110
On social media, sites like Facebook or LinkedIn, regularly recommend friendships.

21
00:01:42,110 --> 00:01:46,610
Recommender systems are even used to personalize your experience on the web.

22
00:01:46,610 --> 00:01:52,229
For example, when you go to a news platform website, a recommender system will make note

23
00:01:52,229 --> 00:01:57,110
of the types of stories that you clicked on and make recommendations on which types of

24
00:01:57,110 --> 00:02:01,120
stories you might be interested in reading, in future.

25
00:02:01,120 --> 00:02:06,270
There are many of these types of examples and they are growing in number every day.

26
00:02:06,270 --> 00:02:12,310
So, let’s take a closer look at the main benefits of using a recommendation system.

27
00:02:12,310 --> 00:02:17,790
One of the main advantages of using recommendation systems is that users get a broader exposure

28
00:02:17,790 --> 00:02:21,650
to many different products they might be interested in.

29
00:02:21,650 --> 00:02:27,930
This exposure encourages users towards continual usage or purchase of their product.

30
00:02:27,930 --> 00:02:33,860
Not only does this provide a better experience for the user but it benefits the service provider,

31
00:02:33,860 --> 00:02:39,930
as well, with increased potential revenue and better security for its customers.

32
00:02:39,930 --> 00:02:46,989
There are generally 2 main types of recommendation systems: Content-based and collaborative filtering.

33
00:02:46,989 --> 00:02:51,680
The main difference between each, can be summed up by the type of statement that a consumer

34
00:02:51,680 --> 00:02:56,590
might make. For instance, the main paradigm of a Content-based

35
00:02:56,590 --> 00:03:02,330
recommendation system is driven by the statement: “Show me more of the same of what I've liked before."

36
00:03:02,330 --> 00:03:06,580
Content-based systems try to figure out what

37
00:03:06,580 --> 00:03:12,770
a user's favorite aspects of an item are, and then make recommendations on items that

38
00:03:12,770 --> 00:03:15,520
share those aspects.

39
00:03:15,520 --> 00:03:20,220
Collaborative filtering is based on a user saying, “Tell me what's popular among my

40
00:03:20,220 --> 00:03:26,099
neighbors because I might like it too.” Collaborative filtering techniques find similar

41
00:03:26,099 --> 00:03:32,110
groups of users, and provide recommendations based on similar tastes within that group.

42
00:03:32,110 --> 00:03:38,209
In short, it assumes that a user might be interested in what similar users are interested in.

43
00:03:38,209 --> 00:03:42,400
Also, there are Hybrid recommender systems,

44
00:03:42,400 --> 00:03:44,510
which combine various mechanisms.

45
00:03:44,510 --> 00:03:53,360
In terms of implementing recommender systems, there are 2 types: Memory-based and Model-based.

46
00:03:53,360 --> 00:04:00,290
In memory-based approaches, we use the entire user-item dataset to generate a recommendation

47
00:04:00,290 --> 00:04:04,120
system. It uses statistical techniques to approximate

48
00:04:04,120 --> 00:04:09,840
users or items. Examples of these techniques include: Pearson

49
00:04:09,840 --> 00:04:15,420
Correlation, Cosine Similarity and Euclidean Distance, among others.

50
00:04:15,420 --> 00:04:20,850
In model-based approaches, a model of users is developed in an attempt to learn their

51
00:04:20,850 --> 00:04:24,860
preferences. Models can be created using Machine Learning

52
00:04:24,860 --> 00:04:30,660
techniques like regression, clustering, classification, and so on.

53
00:04:30,660 --> 00:04:33,349
This is the end of our video. Thanks for watching!