SlideShare una empresa de Scribd logo
1 de 57
Descargar para leer sin conexión
1
DATA MODELING IN IPTV AND OTT RECOMMENDER SYSTEMS –
INFORMATION FLOW FROM INTERACTION TO PERSONALIZED
RECOMMENDATIONS
Dávid Zibriczky, ImpressTV
Neo4j Meetup 2016-02-18
2
Introduction
3
Who we are?
• Acquisition of IPTV, OTT and media business line of Gravity R&D (July 2014)
• Technical Centre in Budapest, sales and management from the UK
What we are doing?
• Providing personalized recommendations, data analytics, targeted advertising and audience
measurement for corporate clients
• Main domains: Video-On-Demand, Linear TV, OTT, Advertisements
About ImpressTV
4
• More time spent watching media contents than ever
• Hundreds of channels, ten thousands of movies, millions of user uploaded videos
• Heterogeneous mixture of devices and items
• Massive consumption information
• Cloudization and centralization
Then and Now
5
Source: http://bigscreenglobal.com/
6
What is a Recommender System?
7
• For the consumers
› Content discovery (relevance, time, information filtering)
› Exploring new preferences (habits, engagement)
• For the business
› Improving KPIs, balancing consumption (long tail contents)
› Promotions, targeting, campaign, analytics, reporting
• For the recommender system vendors
› Data integration/modeling, insights, data science, optimization, research
› Technology challenges, deployment, maintenance
What does a Recommender System mean in TV business?
8
Data modeling
9
Data sources
• Interactions:
› Sources: Remote controller + interface, touchpad, mouse, keyboard
› Explicit / implicit feedbacks: Rating, like, (click) / buy, watch, record, subscribe, channel switch
› Context: Time, date, device, duration, price
• Item catalog:
› Sources: TV provider, external sources (Freebase, DBPedia, IMDb, …)
› Metadata: Descriptive (title, synopsis), technical (schedule, resolution, item type), commercial (price)
• User catalog:
› Sources: Internal, social network
› Metadata: Gender, age group, location
• Recommendation request: user id, item ids, device, location, time, session id
10
Let’s build some graphs…!
Consumers
(Mary)
(Mark)
(John)
Friends S01E01
Friends S02E05
Consumers
Friends S01E02
News 2016-02-
18 07:00
(Mary)
(John)
(Mark)
Contents
Friends S01E01
Friends S02E05
Consumers
Friends S01E02
News 2016-02-
18 07:00
(Mary)
(John)
(Mark)
Contents
Star Wars IV.
VOD (SD)
Star Wars VII.
VOD (HD)
Inception
CatchUp feb 17
Inception
HBO2 22:00
Friends S01E01
SVOD
Inception
PVR 20:00
Inception
HBO 20:00
News
BBC HD 07:00
Items
Inception
(HD)
VOD/SVOD
Linear (EPG items)
CatchUp
Friends S01E01
Friends S02E05
Consumers
Friends S01E02
News 2016-02-
18 07:00
(Mary)
(John)
(Mark)
Contents
Star Wars IV.
VOD (SD)
Star Wars VII.
VOD (HD)
Inception
CatchUp feb 17
Inception
HBO2 22:00
Friends S01E01
SVOD
Inception
PVR 20:00
Inception
HBO 20:00
News
BBC HD 07:00
Items
HBO SD
HBO 2 SD
BBC HD
Inception
(HD)
TV Channels
Friends S01E01
Friends S02E05
Consumers
Friends S01E02
News 2016-02-
18 07:00
(Mary)
(John)
(Mark)
Contents
Star Wars IV.
VOD (SD)
Star Wars VII.
VOD (HD)
Inception
CatchUp feb 17
Inception
HBO2 22:00
Friends S01E01
SVOD
Inception
PVR 20:00
Inception
HBO 20:00
News
BBC HD 07:00
Items
HBO SD
HBO 2 SD
BBC HD
Friends Series
Movie
Channels
Products
Inception
(HD)
Star Wars
Package
Inception
TVOD
($11.99)
($9.99)
($2.99)
VOD/SVOD/Channel products
Friends S01E01
Friends S02E05
Consumers
Friends S01E02
News 2016-02-
18 07:00
(Mary)
(John)
(Mark)
Contents
Star Wars IV.
VOD (SD)
Star Wars VII.
VOD (HD)
Inception
CatchUp feb 17
Inception
HBO2 22:00
Friends S01E01
SVOD
Inception
PVR 20:00
Inception
HBO 20:00
News
BBC HD 07:00
Items
HBO SD
HBO 2 SD
Friends Series
Movie
Channels
Products
Inception
(HD)
Star Wars
Package
($11.99)
($9.99)
($2.99)
Inception
TVOD
Interface
BBC HD
Devices
(remote controller,
mouse, touch pad)
Friends S01E01
Friends S02E05
Consumers
Friends S01E02
News 2016-02-
18 07:00
(Mary)
(John)
(Mark)
Contents
Star Wars IV.
VOD (SD)
Star Wars VII.
VOD (HD)
Inception
CatchUp feb 17
Inception
HBO2 22:00
Friends S01E01
SVOD
Inception
PVR 20:00
Inception
HBO 20:00
News
BBC HD 07:00
Items
HBO SD
HBO 2 SD
Friends Series
Movie
Channels
Products
Inception
(HD)
Star Wars
Package
($11.99)
($9.99)
($2.99)
Inception
TVOD
(19:25; 1136s)
(19:00)
(16:13; 3715s)
(20:10; 1135s)
Interface
BBC HD
(21:11; 3651s; 5*)
Implicit/Explicit
feedbacks + context
Friends S01E01
Friends S02E05
Consumers
Friends S01E02
News 2016-02-
18 07:00
(Mary)
(John)
(Mark)
Contents
Star Wars IV.
VOD (SD)
Star Wars VII.
VOD (HD)
Inception
CatchUp feb 17
Inception
HBO2 22:00
Friends S01E01
SVOD
Inception
PVR 20:00
Inception
HBO 20:00
News
BBC HD 07:00
Items
HBO SD
HBO 2 SD
Friends Series
Movie
Channels
Products
Inception
(HD)
Star Wars
Package
($11.99)
($9.99)
($2.99)
Inception
TVOD
(19:25; 1136s)
(19:00)
(21:11; 3651s; 5*)
(16:13; 3715s)
(20:10; 1135s)
Interface
TV provider related space
BBC HD
Friends S01E01
Friends S02E05
Consumers
Friends S01E02
News 2016-02-
18 07:00
(Mary)
(John)
(Mark)
Contents
Star Wars IV.
VOD (SD)
Star Wars VII.
VOD (HD)
Inception
CatchUp feb 17
Inception
HBO2 22:00
Friends S01E01
SVOD
Inception
PVR 20:00
Inception
HBO 20:00
News
BBC HD 07:00
Items
HBO SD
HBO 2 SD
Friends Series
Movie
Channels
Products
Inception
(HD)
Star Wars
Package
($11.99)
($9.99)
($2.99)
Inception
TVOD
(19:25; 1136s)
(19:00)
(21:11; 3651s; 5*)
(16:13; 3715s)
(20:10; 1135s)
Interface
TV provider related space
TopGear S01
Friends S01
Friends S02
Collections
BBC HD
Friends S01E01
Friends S02E05
Consumers
Friends S01E02
News 2016-02-
18 07:00
(Mary)
(John)
(Mark)
Contents
Star Wars IV.
VOD (SD)
Star Wars VII.
VOD (HD)
Inception
CatchUp feb 17
Inception
HBO2 22:00
Friends S01E01
SVOD
Inception
PVR 20:00
Inception
HBO 20:00
News
BBC HD 07:00
Items
HBO SD
HBO 2 SD
Friends Series
Movie
Channels
Products
Inception
(HD)
Star Wars
Package
($11.99)
($9.99)
($2.99)
Inception
TVOD
(19:25; 1136s)
(19:00)
(21:11; 3651s; 5*)
(16:13; 3715s)
(20:10; 1135s)
Interface
TV provider related space
TopGear S01
Friends S01
Friends S02
Collections
Movies
Series
Shows
Domains
Domain dependent contents
BBC HD
Friends S01E01
Friends S02E05
Consumers
Friends S01E02
News 2016-02-
18 07:00
(Mary)
(John)
(Mark)
Contents
Star Wars IV.
VOD (SD)
Star Wars VII.
VOD (HD)
Inception
CatchUp feb 17
Inception
HBO2 22:00
Friends S01E01
SVOD
Inception
PVR 20:00
Inception
HBO 20:00
News
BBC HD 07:00
Items
HBO SD
HBO 2 SD
Friends Series
Movie
Channels
Products
Inception
(HD)
Star Wars
Package
($11.99)
($9.99)
($2.99)
Inception
TVOD
(19:25; 1136s)
(19:00)
(21:11; 3651s; 5*)
(16:13; 3715s)
(20:10; 1135s)
Interface
TV provider related space
TopGear S01
Friends S01
Friends S02
Collections
Movies
Series
Shows
Domains
TV
Domain dependent contents
Musics
Books
BBC HD
Movies
Series
Shows
Friends S01E01
Friends S02E05
Star Wars IV.
VOD (SD)
Inception
(HD)
Star Wars VII.
VOD (HD)
Inception
CatchUp feb 17
Inception
HBO2 22:00
Friends S01E01
SVOD
Inception
PVR 20:00
Male
Action
Inception
HBO 20:00
Domain dependent contents
Collections Contents Items Products Interface Consumers
TV
Friends S01E02
Domains
TV provider related space
News 2016-02-
18 07:00
News
News
BBC HD 07:00
Harrison
Ford
Female
Comedy
HBO SD
HBO 2 SD
HD
SD
Friends Series
Star Wars
Package
TopGear S01
Friends S01
Friends S02
Musics
(19:25; 1136s)
(19:00)
(16:13; 3715s)
($11.99)
Movie
Channels
($9.99)
(8880s; 20:00-22:28)
(Mary)
(John)
(Mark)
(20:10; 1135s)
Inception
TVOD
($2.99)
(21:11; 3651s; 5*)
BBC HD
Meta data
Books
Movies
Series
Shows
Friends S01E01
Friends S02E05
Star Wars IV.
VOD (SD)
Inception
(HD)
Star Wars VII.
VOD (HD)
Inception
CatchUp feb 17
Inception
HBO2 22:00
Friends S01E01
SVOD
Inception
PVR 20:00
Male
Action
Inception
HBO 20:00
Domain dependent contents
Collections Contents Items Products Interface Consumers
TV
Friends S01E02
Domains
TV provider related space
News 2016-02-
18 07:00
News
News
BBC HD 07:00
Harrison
Ford
Female
Comedy
HBO SD
HBO 2 SD
HD
SD
Friends Series
Star Wars
Package
TopGear S01
Friends S01
Friends S02
Musics
(19:25; 1136s)
(19:00)
(16:13; 3715s)
($11.99)
Movie
Channels
($9.99)
(8880s; 20:00-22:28)
(Mary)
(John)
(Mark)
(20:10; 1135s)
Inception
TVOD
($2.99)
(21:11; 3651s; 5*)
BBC HD
Meta data
Books
24
Graph representation
Nodes
• Entities: Users, items, contents, product, collections
• Metadata: Genres, actors, category tree nodes, gender, age group
• Node properties: Auxiliary meta data (e.g. descriptions)
Edges
• Interactions: Implicit/explicit feedbacks
• Metadata containment, relations
• Edge properties: Context, direction
25
Unification challenges
PVR
LiveTV
CatchUp
SVOD
TVOD
ADs
Devices
Social Networks
Metadata sources
TV items (providers)
26
Recommendations
27
Evolution of the Recommender Problem
Source: http://www.slideshare.net/xamat/recsys-2014-tutorial-the-recommender-problem-revisited
28
29
Collaborative Filtering
Movies
Series
Shows
Friends S01E01
Friends S02E05
Star Wars IV.
VOD (SD)
Inception
(HD)
Star Wars VII.
VOD (HD)
Inception
CatchUp feb 17
Inception
HBO2 22:00
Friends S01E01
SVOD
Inception
PVR 20:00
Male
Action
Inception
HBO 20:00
Domain dependent contents
Collections Contents Items Products Interface Consumers
TV
Friends S01E02
Domains
TV provider related space
News 2016-02-
18 07:00
News
News
BBC HD 07:00
Harrison
Ford
Female
Comedy
HBO SD
HBO 2 SD
HD
SD
Friends Series
Star Wars
Package
TopGear S01
Friends S01
Friends S02
Musics
(19:25; 1136s)
(19:00)
(16:13; 3715s)
($11.99)
Movie
Channels
($9.99)
(8880s; 20:00-22:28)
(Mary)
(John)
(Mark)
(20:10; 1135s)
Inception
TVOD
($2.99)
(21:11; 3651s; 5*)
BBC HD
Meta data
Books
Movies
Series
Shows
Friends S01E01
Friends S02E05
Star Wars IV.
VOD (SD)
Inception
(HD)
Star Wars VII.
VOD (HD)
Inception
CatchUp feb 17
Inception
HBO2 22:00
Friends S01E01
SVOD
Inception
PVR 20:00
Male
Action
Inception
HBO 20:00
Domain dependent contents
Collections Contents Items Products Interface Consumers
TV
Friends S01E02
Domains
TV provider related space
News 2016-02-
18 07:00
News
News
BBC HD 07:00
Harrison
Ford
Female
Comedy
HBO SD
HBO 2 SD
HD
SD
Friends Series
Star Wars
Package
TopGear S01
Friends S01
Friends S02
Musics
(19:25; 1136s)
(19:00)
(16:13; 3715s)
($11.99)
Movie
Channels
($9.99)
(8880s; 20:00-22:28)
(Mary)
(John)
(Mark)
(20:10; 1135s)
Inception
TVOD
($2.99)
(21:11; 3651s; 5*)
BBC HD
Meta data
Books
Star Wars IV.
VOD (SD)
Inception
(HD)
Star Wars VII.
VOD (HD)
Inception
CatchUp feb 17
Inception
HBO2 22:00
Friends S01E01
SVOD
Inception
PVR 20:00
Inception
HBO 20:00
Items Products Interface Consumers
News
BBC HD 07:00
HBO SD
HBO 2 SD
Friends Series
Star Wars
Package
(19:25; 1136s)
(19:00)
(16:13; 3715s)
($11.99)
Movie
Channels
($9.99)
(8880s; 20:00-22:28)
(Mary)
(John)
(Mark)
(20:10; 1135s)
Inception
TVOD
($2.99)
(21:11; 3651s; 5*)
BBC HD
Star Wars IV.
VOD (SD)
Inception
(HD)
Star Wars VII.
VOD (HD)
Inception
CatchUp feb 17
Inception
HBO2 22:00
Friends S01E01
SVOD
Inception
PVR 20:00
Inception
HBO 20:00
Items Consumers
News
BBC HD 07:00
(8880s; 20:00-22:28)
(Mary)
(John)
(Mark)
34
Collaborative Filtering / Graph representation
The
Matrix
The
Matrix 2
The
Matrix 3
Twilight
35
Collaborative Filtering / Matrix representation
The
Matrix
The
Matrix 2
The
Matrix 3
Twilight
R
The
Matrix
The
Matrix 2
Twilight
The
Matrix 3
User 1 1 1 ? ?
User 2 1 1 0 0
User 3 1 0 1 1
User 4 1 0 0 1
?
?
36
Collaborative Filtering / Binary preference matrix
R
The
Matrix
The
Matrix 2
Twilight
The
Matrix 3
User 1 1 1 ? ?
User 2 1 1 0 0
User 3 1 0 1 1
User 4 1 0 0 1
37
Collaborative Filtering / Implicit MF
R
The
Matrix
The
Matrix 2
Twilight
The
Matrix 3
User 1
User 2 𝒓 𝑢𝑖
User 3
User 4
Q 𝒒𝑖
PT
𝒑 𝑢
𝑇
𝑹 𝑵𝒙𝑴: preference matrix
𝑷 𝑵𝒙𝑲: user feature matrix
𝑸 𝑴𝒙𝑲: item feature matrix
𝑵: #users
𝑴: #items
𝑲: #features
𝑲 ≪ 𝑴 , 𝑲 ≪ 𝑵
𝑹~𝑷 𝐓 𝑸 𝑟 𝑢𝑖 = 𝒑 𝑢
T 𝒒𝑖
38
Collaborative Filtering / Implicit MF / Weighted Sum of Squared Errors
R
The
Matrix
The
Matrix 2
Twilight
The
Matrix 3
User 1 1 1 ? ?
User 2 1 1 0 0
User 3 1 0 1 1
User 4 1 0 0 1
C
The
Matrix
The
Matrix 2
Twilight
The
Matrix 3
User 1 c11 c12 1 1
User 2 c21 c22 1 1
User 3 c31 1 c33 c34
User 4 c41 1 1 1
𝒇 𝑷, 𝑸 = 𝑾𝑺𝑺𝑬 =
(𝒖,𝒊)
𝒄 𝒖𝒊 𝒓 𝒖𝒊 − 𝒓 𝒖𝒊
𝟐• Objective Function:
39
Collaborative Filtering / Implicit MF / Alternating Least Squares
R
The
Matrix
The
Matrix 2
Twilight
The
Matrix 3
User 1
User 2
User 3
User 4
Q
0.1 -0.4 0.8 0.3
0.6 0.7 -0.1 0.8
PT
-0.2 0.6
0.6 0.4
0.7 -0.2
0.5 -0.2
𝑹~𝑷 𝐓 𝑸 𝑟 𝑢𝑖 = 𝒑 𝑢
T 𝒒𝑖
Ridge Regression
• 𝑝 𝑢 = 𝑄𝐶 𝑢
𝑄T −1
𝑄𝐶 𝑢
𝑅 𝑟 𝑢
• 𝑞𝑖 = 𝑃𝐶 𝑖 𝑃T −1
𝑃𝐶 𝑖 𝑅 𝑐 𝑖
40
Collaborative Filtering / Implicit MF / Alternating Least Squares
R
The
Matrix
The
Matrix 2
Twilight
The
Matrix 3
User 1
User 2
User 3
User 4
Q
0.3 -0.3 0.6 0.2
0.7 0.8 0.2 0.7
PT
-0.2 0.6
0.6 0.4
0.7 -0.2
0.5 -0.2
Ridge Regression
• 𝑝 𝑢 = 𝑄𝐶 𝑢
𝑄T −1
𝑄𝐶 𝑢
𝑅 𝑟 𝑢
• 𝑞𝑖 = 𝑃𝐶 𝑖 𝑃T −1
𝑃𝐶 𝑖 𝑅 𝑐 𝑖
𝑹~𝑷 𝐓 𝑸 𝑟 𝑢𝑖 = 𝒑 𝑢
T 𝒒𝑖
41
Collaborative Filtering / Implicit MF / Alternating Least Squares
R
The
Matrix
The
Matrix 2
Twilight
The
Matrix 3
User 1
User 2
User 3
User 4
Q
0.3 -0.3 0.6 0.2
0.7 0.8 0.2 0.7
PT
-0.2 0.7
0.6 0.5
0.8 -0.2
0.6 -0.2
Ridge Regression
• 𝑝 𝑢 = 𝑄𝐶 𝑢
𝑄T −1
𝑄𝐶 𝑢
𝑅 𝑟 𝑢
• 𝑞𝑖 = 𝑃𝐶 𝑖 𝑃T −1
𝑃𝐶 𝑖 𝑅 𝑐 𝑖
𝑹~𝑷 𝐓 𝑸 𝑟 𝑢𝑖 = 𝒑 𝑢
T 𝒒𝑖
42
Collaborative Filtering / Implicit MF / Alternating Least Squares
R
The
Matrix
The
Matrix 2
Twilight
The
Matrix 3
User 1 0.02 0.45
User 2
User 3
User 4
Q
0.3 -0.3 0.6 0.2
0.7 0.8 0.2 0.7
PT
-0.2 0.7
0.6 0.5
0.8 -0.2
0.6 -0.2
𝑹~𝑷 𝐓 𝑸 𝑟 𝑢𝑖 = 𝒑 𝑢
T 𝒒𝑖
43
Content-based Filtering
Movies
Series
Shows
Friends S01E01
Friends S02E05
Star Wars IV.
VOD (SD)
Inception
(HD)
Star Wars VII.
VOD (HD)
Inception
CatchUp feb 17
Inception
HBO2 22:00
Friends S01E01
SVOD
Inception
PVR 20:00
Male
Action
Inception
HBO 20:00
Domain dependent contents
Collections Contents Items Products Interface Consumers
TV
Friends S01E02
Domains
TV provider related space
News 2016-02-
18 07:00
News
News
BBC HD 07:00
Harrison
Ford
Female
Comedy
HBO SD
HBO 2 SD
HD
SD
Friends Series
Star Wars
Package
TopGear S01
Friends S01
Friends S02
Musics
(19:25; 1136s)
(19:00)
(16:13; 3715s)
($11.99)
Movie
Channels
($9.99)
(8880s; 20:00-22:28)
(Mary)
(John)
(Mark)
(20:10; 1135s)
Inception
TVOD
($2.99)
(21:11; 3651s; 5*)
BBC HD
Meta data
Books
Star Wars IV.
VOD (SD)
Inception
(HD)
Star Wars VII.
VOD (HD)
Inception
CatchUp feb 17
Inception
HBO2 22:00
Friends S01E01
SVOD
Inception
PVR 20:00
Action
Inception
HBO 20:00
Contents Items Products Interface Consumers
News
News
BBC HD 07:00
Harrison
Ford
Comedy
HBO SD
HBO 2 SD
HD
SD
Friends Series
Star Wars
Package
(16:13; 3715s)
Movie
Channels
(8880s; 20:00-22:28)
(John)
(20:10; 1135s)
Inception
TVOD
BBC HD
Star Wars IV.
VOD (SD)
Inception
(HD)
Star Wars VII.
VOD (HD)
Inception
CatchUp feb 17
Inception
HBO2 22:00
Friends S01E01
SVOD
Inception
PVR 20:00
Action
Inception
HBO 20:00
Contents Items Products Consumers
News
News
BBC HD 07:00
Harrison
Ford
Comedy
HD
SD
(8880s; 20:00-22:28)
(John)
47
Content-based Filtering / Cosine Similarity (1/6)
Sci-Fi RomanceAdventure
The
Matrix
The
Matrix 2
The
Matrix 3
Twilight
48
Content-based Filtering / Cosine Similarity (2/6)
Sci-Fi RomanceAdventure
The
Matrix
The
Matrix 2
The
Matrix 3
Twilight
?
?
MT Sci-Fi Adventure Romance
The
Matrix
1 1 0
The
Matrix 2
1 1 0
Twilight 0 1 1
The
Matrix 3
1 1 0
R
The
Matrix
The
Matrix 2
Twilight
The
Matrix 3
User 1 1 ? ?
49
Content-based Filtering / Cosine Similarity (3/6)
MT Sci-Fi Adventure Romance
The
Matrix
1 1 0
The
Matrix 2
1 1 0
Twilight 0 1 1
The
Matrix 3
1 1 0
R
The
Matrix
The
Matrix 2
Twilight
The
Matrix 3
User 1 1 ? ?
50
Content-based Filtering / Cosine Similarity (4/6)
M’T Sci-Fi Adventure Romance
The
Matrix
1/6 1/8 0
The
Matrix 2
1/6 1/8 0
Twilight 0 1/8 1/2
The
Matrix 3
1/6 1/8 0
R
The
Matrix
The
Matrix 2
Twilight
The
Matrix 3
User 1 1 ? ?
𝑚′𝑖𝑗 =
𝑚𝑖𝑗
𝑘 𝑚𝑖𝑘 𝑘 𝑚 𝑘𝑗
51
Content-based Filtering / Cosine Similarity (5/6)
M’T Sci-Fi Adventure Romance
The
Matrix
1/6 1/8 0
The
Matrix 2
1/6 1/8 0
Twilight 0 1/8 1/2
The
Matrix 3
1/6 1/8 0
R
The
Matrix
The
Matrix 2
Twilight
The
Matrix 3
User 1 1 ? ?
PT Sci-Fi Adventure Romance
User 2/6 2/8 0
𝑃T = 𝑅𝑀′T
52
Content-based Filtering / Cosine Similarity (6/6)
M’T Sci-Fi Adventure Romance
The
Matrix
1/6 1/8 0
The
Matrix 2
1/6 1/8 0
Twilight 0 1/8 1/2
The
Matrix 3
1/6 1/8 0
R
The
Matrix
The
Matrix 2
Twilight
The
Matrix 3
User 1 1 ? ?
PT Sci-Fi Adventure Romance
User 2/6 2/8 0
𝑹
The
Matrix
The
Matrix 2
Twilight
The
Matrix 3
User 1 1 0.268 1
𝑟𝑢𝑖 =
𝑘 𝑚′𝑖𝑘 𝑝 𝑢𝑘
𝑘 𝑚′𝑖𝑘
2
𝑘 𝑝 𝑢𝑘
2
𝑃T = 𝑅𝑀′T
53
Challenges and research
54
• Who is the front of the TV? (watching behavior pattern recognition)
• Cold-start problem, opt-out users (metadata enrichment, hybrid filtering)
• Noise filtering, lack of explicit data (filtering and weighting method, implicit feedbacks)
• Context-dependency (device, time, mood)
• Cross-domain recommendation, content centralization
• Time-dependent recommendable set (model-based methods)
• Popularity effect (diversification)
• Upselling, subscription (Live  VOD)
• Offline vs. Online KPIs
Data Science Challenges
55
• Heterogeneous external data sources (data unification)
• Big data (scalable algorithms, distributed databases)
• Latency in data transfer, streaming service
• Responsivity, load, SLA (distributed systems, load balancer)
• Maintenance, service, follow-up
• Open source vs. self-development
• Distributed systems vs. single server
• Software-as-a-service vs. on-site deployment
Technology Challenges
56
Presented by:
Contact:
THANK YOU!
www.impresstv.com
David Zibriczky
Head of Data Science
david.zibriczky@impresstv.com
57
• Senior Software Engineers (Java, .NET)
• Developer (Java, Python, Node.js)
• Senior Data Mining Engineer
• Pre-sales Technical Consultant
• Data Scientist
• Contact: hr@impresstv.com
• Website: www.impresstv.com
We are hiring!

Más contenido relacionado

Similar a Data Modeling in IPTV and OTT Recommender Systems

Babcock Media Services
Babcock Media ServicesBabcock Media Services
Babcock Media Servicesbabcockmedia
 
Great IPTV Subs in.pdf
Great IPTV Subs in.pdfGreat IPTV Subs in.pdf
Great IPTV Subs in.pdfgreatiptvsubs
 
Australian Online Feature Films - Stuart Cunningham SPAA 2010
Australian Online Feature Films - Stuart Cunningham SPAA 2010Australian Online Feature Films - Stuart Cunningham SPAA 2010
Australian Online Feature Films - Stuart Cunningham SPAA 2010simonbritton88
 

Similar a Data Modeling in IPTV and OTT Recommender Systems (6)

Babcock Media Services
Babcock Media ServicesBabcock Media Services
Babcock Media Services
 
OTT Service Providers.pdf
OTT Service Providers.pdfOTT Service Providers.pdf
OTT Service Providers.pdf
 
Great IPTV Subs.pdf
Great IPTV Subs.pdfGreat IPTV Subs.pdf
Great IPTV Subs.pdf
 
Great IPTV Subs.pdf
Great IPTV Subs.pdfGreat IPTV Subs.pdf
Great IPTV Subs.pdf
 
Great IPTV Subs in.pdf
Great IPTV Subs in.pdfGreat IPTV Subs in.pdf
Great IPTV Subs in.pdf
 
Australian Online Feature Films - Stuart Cunningham SPAA 2010
Australian Online Feature Films - Stuart Cunningham SPAA 2010Australian Online Feature Films - Stuart Cunningham SPAA 2010
Australian Online Feature Films - Stuart Cunningham SPAA 2010
 

Más de David Zibriczky

Highlights from the 8th ACM Conference on Recommender Systems (RecSys 2014)
Highlights from the 8th ACM Conference on Recommender Systems (RecSys 2014)Highlights from the 8th ACM Conference on Recommender Systems (RecSys 2014)
Highlights from the 8th ACM Conference on Recommender Systems (RecSys 2014)David Zibriczky
 
Predictive Solutions and Analytics for TV & Entertainment Businesses
Predictive Solutions and Analytics for TV & Entertainment BusinessesPredictive Solutions and Analytics for TV & Entertainment Businesses
Predictive Solutions and Analytics for TV & Entertainment BusinessesDavid Zibriczky
 
Improving the TV User Experience by Algorithms: Personalized Content Recommen...
Improving the TV User Experience by Algorithms: Personalized Content Recommen...Improving the TV User Experience by Algorithms: Personalized Content Recommen...
Improving the TV User Experience by Algorithms: Personalized Content Recommen...David Zibriczky
 
Recommender Systems meet Finance - A literature review
Recommender Systems meet Finance - A literature reviewRecommender Systems meet Finance - A literature review
Recommender Systems meet Finance - A literature reviewDavid Zibriczky
 
A Combination of Simple Models by Forward Predictor Selection for Job Recomme...
A Combination of Simple Models by Forward Predictor Selection for Job Recomme...A Combination of Simple Models by Forward Predictor Selection for Job Recomme...
A Combination of Simple Models by Forward Predictor Selection for Job Recomme...David Zibriczky
 
Fast ALS-Based Matrix Factorization for Recommender Systems
Fast ALS-Based Matrix Factorization for Recommender SystemsFast ALS-Based Matrix Factorization for Recommender Systems
Fast ALS-Based Matrix Factorization for Recommender SystemsDavid Zibriczky
 
EPG content recommendation in large scale: a case study on interactive TV pla...
EPG content recommendation in large scale: a case study on interactive TV pla...EPG content recommendation in large scale: a case study on interactive TV pla...
EPG content recommendation in large scale: a case study on interactive TV pla...David Zibriczky
 
Personalized recommendation of linear content on interactive TV platforms
Personalized recommendation of linear content on interactive TV platformsPersonalized recommendation of linear content on interactive TV platforms
Personalized recommendation of linear content on interactive TV platformsDavid Zibriczky
 
An introduction to Recommender Systems
An introduction to Recommender SystemsAn introduction to Recommender Systems
An introduction to Recommender SystemsDavid Zibriczky
 
Entropy based asset pricing
Entropy based asset pricingEntropy based asset pricing
Entropy based asset pricingDavid Zibriczky
 

Más de David Zibriczky (10)

Highlights from the 8th ACM Conference on Recommender Systems (RecSys 2014)
Highlights from the 8th ACM Conference on Recommender Systems (RecSys 2014)Highlights from the 8th ACM Conference on Recommender Systems (RecSys 2014)
Highlights from the 8th ACM Conference on Recommender Systems (RecSys 2014)
 
Predictive Solutions and Analytics for TV & Entertainment Businesses
Predictive Solutions and Analytics for TV & Entertainment BusinessesPredictive Solutions and Analytics for TV & Entertainment Businesses
Predictive Solutions and Analytics for TV & Entertainment Businesses
 
Improving the TV User Experience by Algorithms: Personalized Content Recommen...
Improving the TV User Experience by Algorithms: Personalized Content Recommen...Improving the TV User Experience by Algorithms: Personalized Content Recommen...
Improving the TV User Experience by Algorithms: Personalized Content Recommen...
 
Recommender Systems meet Finance - A literature review
Recommender Systems meet Finance - A literature reviewRecommender Systems meet Finance - A literature review
Recommender Systems meet Finance - A literature review
 
A Combination of Simple Models by Forward Predictor Selection for Job Recomme...
A Combination of Simple Models by Forward Predictor Selection for Job Recomme...A Combination of Simple Models by Forward Predictor Selection for Job Recomme...
A Combination of Simple Models by Forward Predictor Selection for Job Recomme...
 
Fast ALS-Based Matrix Factorization for Recommender Systems
Fast ALS-Based Matrix Factorization for Recommender SystemsFast ALS-Based Matrix Factorization for Recommender Systems
Fast ALS-Based Matrix Factorization for Recommender Systems
 
EPG content recommendation in large scale: a case study on interactive TV pla...
EPG content recommendation in large scale: a case study on interactive TV pla...EPG content recommendation in large scale: a case study on interactive TV pla...
EPG content recommendation in large scale: a case study on interactive TV pla...
 
Personalized recommendation of linear content on interactive TV platforms
Personalized recommendation of linear content on interactive TV platformsPersonalized recommendation of linear content on interactive TV platforms
Personalized recommendation of linear content on interactive TV platforms
 
An introduction to Recommender Systems
An introduction to Recommender SystemsAn introduction to Recommender Systems
An introduction to Recommender Systems
 
Entropy based asset pricing
Entropy based asset pricingEntropy based asset pricing
Entropy based asset pricing
 

Último

Unveiling the Role of Social Media Suspect Investigators in Preventing Online...
Unveiling the Role of Social Media Suspect Investigators in Preventing Online...Unveiling the Role of Social Media Suspect Investigators in Preventing Online...
Unveiling the Role of Social Media Suspect Investigators in Preventing Online...Milind Agarwal
 
Networking Case Study prepared by teacher.pptx
Networking Case Study prepared by teacher.pptxNetworking Case Study prepared by teacher.pptx
Networking Case Study prepared by teacher.pptxHimangsuNath
 
Decoding Patterns: Customer Churn Prediction Data Analysis Project
Decoding Patterns: Customer Churn Prediction Data Analysis ProjectDecoding Patterns: Customer Churn Prediction Data Analysis Project
Decoding Patterns: Customer Churn Prediction Data Analysis ProjectBoston Institute of Analytics
 
Data Analysis Project Presentation: Unveiling Your Ideal Customer, Bank Custo...
Data Analysis Project Presentation: Unveiling Your Ideal Customer, Bank Custo...Data Analysis Project Presentation: Unveiling Your Ideal Customer, Bank Custo...
Data Analysis Project Presentation: Unveiling Your Ideal Customer, Bank Custo...Boston Institute of Analytics
 
Minimizing AI Hallucinations/Confabulations and the Path towards AGI with Exa...
Minimizing AI Hallucinations/Confabulations and the Path towards AGI with Exa...Minimizing AI Hallucinations/Confabulations and the Path towards AGI with Exa...
Minimizing AI Hallucinations/Confabulations and the Path towards AGI with Exa...Thomas Poetter
 
Digital Marketing Plan, how digital marketing works
Digital Marketing Plan, how digital marketing worksDigital Marketing Plan, how digital marketing works
Digital Marketing Plan, how digital marketing worksdeepakthakur548787
 
Student profile product demonstration on grades, ability, well-being and mind...
Student profile product demonstration on grades, ability, well-being and mind...Student profile product demonstration on grades, ability, well-being and mind...
Student profile product demonstration on grades, ability, well-being and mind...Seán Kennedy
 
Advanced Machine Learning for Business Professionals
Advanced Machine Learning for Business ProfessionalsAdvanced Machine Learning for Business Professionals
Advanced Machine Learning for Business ProfessionalsVICTOR MAESTRE RAMIREZ
 
Cyber awareness ppt on the recorded data
Cyber awareness ppt on the recorded dataCyber awareness ppt on the recorded data
Cyber awareness ppt on the recorded dataTecnoIncentive
 
Defining Constituents, Data Vizzes and Telling a Data Story
Defining Constituents, Data Vizzes and Telling a Data StoryDefining Constituents, Data Vizzes and Telling a Data Story
Defining Constituents, Data Vizzes and Telling a Data StoryJeremy Anderson
 
Data Factory in Microsoft Fabric (MsBIP #82)
Data Factory in Microsoft Fabric (MsBIP #82)Data Factory in Microsoft Fabric (MsBIP #82)
Data Factory in Microsoft Fabric (MsBIP #82)Cathrine Wilhelmsen
 
What To Do For World Nature Conservation Day by Slidesgo.pptx
What To Do For World Nature Conservation Day by Slidesgo.pptxWhat To Do For World Nature Conservation Day by Slidesgo.pptx
What To Do For World Nature Conservation Day by Slidesgo.pptxSimranPal17
 
Student Profile Sample report on improving academic performance by uniting gr...
Student Profile Sample report on improving academic performance by uniting gr...Student Profile Sample report on improving academic performance by uniting gr...
Student Profile Sample report on improving academic performance by uniting gr...Seán Kennedy
 
Predictive Analysis for Loan Default Presentation : Data Analysis Project PPT
Predictive Analysis for Loan Default  Presentation : Data Analysis Project PPTPredictive Analysis for Loan Default  Presentation : Data Analysis Project PPT
Predictive Analysis for Loan Default Presentation : Data Analysis Project PPTBoston Institute of Analytics
 
INTRODUCTION TO Natural language processing
INTRODUCTION TO Natural language processingINTRODUCTION TO Natural language processing
INTRODUCTION TO Natural language processingsocarem879
 
English-8-Q4-W3-Synthesizing-Essential-Information-From-Various-Sources-1.pdf
English-8-Q4-W3-Synthesizing-Essential-Information-From-Various-Sources-1.pdfEnglish-8-Q4-W3-Synthesizing-Essential-Information-From-Various-Sources-1.pdf
English-8-Q4-W3-Synthesizing-Essential-Information-From-Various-Sources-1.pdfblazblazml
 
Semantic Shed - Squashing and Squeezing.pptx
Semantic Shed - Squashing and Squeezing.pptxSemantic Shed - Squashing and Squeezing.pptx
Semantic Shed - Squashing and Squeezing.pptxMike Bennett
 
Principles and Practices of Data Visualization
Principles and Practices of Data VisualizationPrinciples and Practices of Data Visualization
Principles and Practices of Data VisualizationKianJazayeri1
 
Conf42-LLM_Adding Generative AI to Real-Time Streaming Pipelines
Conf42-LLM_Adding Generative AI to Real-Time Streaming PipelinesConf42-LLM_Adding Generative AI to Real-Time Streaming Pipelines
Conf42-LLM_Adding Generative AI to Real-Time Streaming PipelinesTimothy Spann
 
Data Analysis Project : Targeting the Right Customers, Presentation on Bank M...
Data Analysis Project : Targeting the Right Customers, Presentation on Bank M...Data Analysis Project : Targeting the Right Customers, Presentation on Bank M...
Data Analysis Project : Targeting the Right Customers, Presentation on Bank M...Boston Institute of Analytics
 

Último (20)

Unveiling the Role of Social Media Suspect Investigators in Preventing Online...
Unveiling the Role of Social Media Suspect Investigators in Preventing Online...Unveiling the Role of Social Media Suspect Investigators in Preventing Online...
Unveiling the Role of Social Media Suspect Investigators in Preventing Online...
 
Networking Case Study prepared by teacher.pptx
Networking Case Study prepared by teacher.pptxNetworking Case Study prepared by teacher.pptx
Networking Case Study prepared by teacher.pptx
 
Decoding Patterns: Customer Churn Prediction Data Analysis Project
Decoding Patterns: Customer Churn Prediction Data Analysis ProjectDecoding Patterns: Customer Churn Prediction Data Analysis Project
Decoding Patterns: Customer Churn Prediction Data Analysis Project
 
Data Analysis Project Presentation: Unveiling Your Ideal Customer, Bank Custo...
Data Analysis Project Presentation: Unveiling Your Ideal Customer, Bank Custo...Data Analysis Project Presentation: Unveiling Your Ideal Customer, Bank Custo...
Data Analysis Project Presentation: Unveiling Your Ideal Customer, Bank Custo...
 
Minimizing AI Hallucinations/Confabulations and the Path towards AGI with Exa...
Minimizing AI Hallucinations/Confabulations and the Path towards AGI with Exa...Minimizing AI Hallucinations/Confabulations and the Path towards AGI with Exa...
Minimizing AI Hallucinations/Confabulations and the Path towards AGI with Exa...
 
Digital Marketing Plan, how digital marketing works
Digital Marketing Plan, how digital marketing worksDigital Marketing Plan, how digital marketing works
Digital Marketing Plan, how digital marketing works
 
Student profile product demonstration on grades, ability, well-being and mind...
Student profile product demonstration on grades, ability, well-being and mind...Student profile product demonstration on grades, ability, well-being and mind...
Student profile product demonstration on grades, ability, well-being and mind...
 
Advanced Machine Learning for Business Professionals
Advanced Machine Learning for Business ProfessionalsAdvanced Machine Learning for Business Professionals
Advanced Machine Learning for Business Professionals
 
Cyber awareness ppt on the recorded data
Cyber awareness ppt on the recorded dataCyber awareness ppt on the recorded data
Cyber awareness ppt on the recorded data
 
Defining Constituents, Data Vizzes and Telling a Data Story
Defining Constituents, Data Vizzes and Telling a Data StoryDefining Constituents, Data Vizzes and Telling a Data Story
Defining Constituents, Data Vizzes and Telling a Data Story
 
Data Factory in Microsoft Fabric (MsBIP #82)
Data Factory in Microsoft Fabric (MsBIP #82)Data Factory in Microsoft Fabric (MsBIP #82)
Data Factory in Microsoft Fabric (MsBIP #82)
 
What To Do For World Nature Conservation Day by Slidesgo.pptx
What To Do For World Nature Conservation Day by Slidesgo.pptxWhat To Do For World Nature Conservation Day by Slidesgo.pptx
What To Do For World Nature Conservation Day by Slidesgo.pptx
 
Student Profile Sample report on improving academic performance by uniting gr...
Student Profile Sample report on improving academic performance by uniting gr...Student Profile Sample report on improving academic performance by uniting gr...
Student Profile Sample report on improving academic performance by uniting gr...
 
Predictive Analysis for Loan Default Presentation : Data Analysis Project PPT
Predictive Analysis for Loan Default  Presentation : Data Analysis Project PPTPredictive Analysis for Loan Default  Presentation : Data Analysis Project PPT
Predictive Analysis for Loan Default Presentation : Data Analysis Project PPT
 
INTRODUCTION TO Natural language processing
INTRODUCTION TO Natural language processingINTRODUCTION TO Natural language processing
INTRODUCTION TO Natural language processing
 
English-8-Q4-W3-Synthesizing-Essential-Information-From-Various-Sources-1.pdf
English-8-Q4-W3-Synthesizing-Essential-Information-From-Various-Sources-1.pdfEnglish-8-Q4-W3-Synthesizing-Essential-Information-From-Various-Sources-1.pdf
English-8-Q4-W3-Synthesizing-Essential-Information-From-Various-Sources-1.pdf
 
Semantic Shed - Squashing and Squeezing.pptx
Semantic Shed - Squashing and Squeezing.pptxSemantic Shed - Squashing and Squeezing.pptx
Semantic Shed - Squashing and Squeezing.pptx
 
Principles and Practices of Data Visualization
Principles and Practices of Data VisualizationPrinciples and Practices of Data Visualization
Principles and Practices of Data Visualization
 
Conf42-LLM_Adding Generative AI to Real-Time Streaming Pipelines
Conf42-LLM_Adding Generative AI to Real-Time Streaming PipelinesConf42-LLM_Adding Generative AI to Real-Time Streaming Pipelines
Conf42-LLM_Adding Generative AI to Real-Time Streaming Pipelines
 
Data Analysis Project : Targeting the Right Customers, Presentation on Bank M...
Data Analysis Project : Targeting the Right Customers, Presentation on Bank M...Data Analysis Project : Targeting the Right Customers, Presentation on Bank M...
Data Analysis Project : Targeting the Right Customers, Presentation on Bank M...
 

Data Modeling in IPTV and OTT Recommender Systems

  • 1. 1 DATA MODELING IN IPTV AND OTT RECOMMENDER SYSTEMS – INFORMATION FLOW FROM INTERACTION TO PERSONALIZED RECOMMENDATIONS Dávid Zibriczky, ImpressTV Neo4j Meetup 2016-02-18
  • 3. 3 Who we are? • Acquisition of IPTV, OTT and media business line of Gravity R&D (July 2014) • Technical Centre in Budapest, sales and management from the UK What we are doing? • Providing personalized recommendations, data analytics, targeted advertising and audience measurement for corporate clients • Main domains: Video-On-Demand, Linear TV, OTT, Advertisements About ImpressTV
  • 4. 4 • More time spent watching media contents than ever • Hundreds of channels, ten thousands of movies, millions of user uploaded videos • Heterogeneous mixture of devices and items • Massive consumption information • Cloudization and centralization Then and Now
  • 6. 6 What is a Recommender System?
  • 7. 7 • For the consumers › Content discovery (relevance, time, information filtering) › Exploring new preferences (habits, engagement) • For the business › Improving KPIs, balancing consumption (long tail contents) › Promotions, targeting, campaign, analytics, reporting • For the recommender system vendors › Data integration/modeling, insights, data science, optimization, research › Technology challenges, deployment, maintenance What does a Recommender System mean in TV business?
  • 9. 9 Data sources • Interactions: › Sources: Remote controller + interface, touchpad, mouse, keyboard › Explicit / implicit feedbacks: Rating, like, (click) / buy, watch, record, subscribe, channel switch › Context: Time, date, device, duration, price • Item catalog: › Sources: TV provider, external sources (Freebase, DBPedia, IMDb, …) › Metadata: Descriptive (title, synopsis), technical (schedule, resolution, item type), commercial (price) • User catalog: › Sources: Internal, social network › Metadata: Gender, age group, location • Recommendation request: user id, item ids, device, location, time, session id
  • 10. 10 Let’s build some graphs…!
  • 12. Friends S01E01 Friends S02E05 Consumers Friends S01E02 News 2016-02- 18 07:00 (Mary) (John) (Mark) Contents
  • 13. Friends S01E01 Friends S02E05 Consumers Friends S01E02 News 2016-02- 18 07:00 (Mary) (John) (Mark) Contents Star Wars IV. VOD (SD) Star Wars VII. VOD (HD) Inception CatchUp feb 17 Inception HBO2 22:00 Friends S01E01 SVOD Inception PVR 20:00 Inception HBO 20:00 News BBC HD 07:00 Items Inception (HD) VOD/SVOD Linear (EPG items) CatchUp
  • 14. Friends S01E01 Friends S02E05 Consumers Friends S01E02 News 2016-02- 18 07:00 (Mary) (John) (Mark) Contents Star Wars IV. VOD (SD) Star Wars VII. VOD (HD) Inception CatchUp feb 17 Inception HBO2 22:00 Friends S01E01 SVOD Inception PVR 20:00 Inception HBO 20:00 News BBC HD 07:00 Items HBO SD HBO 2 SD BBC HD Inception (HD) TV Channels
  • 15. Friends S01E01 Friends S02E05 Consumers Friends S01E02 News 2016-02- 18 07:00 (Mary) (John) (Mark) Contents Star Wars IV. VOD (SD) Star Wars VII. VOD (HD) Inception CatchUp feb 17 Inception HBO2 22:00 Friends S01E01 SVOD Inception PVR 20:00 Inception HBO 20:00 News BBC HD 07:00 Items HBO SD HBO 2 SD BBC HD Friends Series Movie Channels Products Inception (HD) Star Wars Package Inception TVOD ($11.99) ($9.99) ($2.99) VOD/SVOD/Channel products
  • 16. Friends S01E01 Friends S02E05 Consumers Friends S01E02 News 2016-02- 18 07:00 (Mary) (John) (Mark) Contents Star Wars IV. VOD (SD) Star Wars VII. VOD (HD) Inception CatchUp feb 17 Inception HBO2 22:00 Friends S01E01 SVOD Inception PVR 20:00 Inception HBO 20:00 News BBC HD 07:00 Items HBO SD HBO 2 SD Friends Series Movie Channels Products Inception (HD) Star Wars Package ($11.99) ($9.99) ($2.99) Inception TVOD Interface BBC HD Devices (remote controller, mouse, touch pad)
  • 17. Friends S01E01 Friends S02E05 Consumers Friends S01E02 News 2016-02- 18 07:00 (Mary) (John) (Mark) Contents Star Wars IV. VOD (SD) Star Wars VII. VOD (HD) Inception CatchUp feb 17 Inception HBO2 22:00 Friends S01E01 SVOD Inception PVR 20:00 Inception HBO 20:00 News BBC HD 07:00 Items HBO SD HBO 2 SD Friends Series Movie Channels Products Inception (HD) Star Wars Package ($11.99) ($9.99) ($2.99) Inception TVOD (19:25; 1136s) (19:00) (16:13; 3715s) (20:10; 1135s) Interface BBC HD (21:11; 3651s; 5*) Implicit/Explicit feedbacks + context
  • 18. Friends S01E01 Friends S02E05 Consumers Friends S01E02 News 2016-02- 18 07:00 (Mary) (John) (Mark) Contents Star Wars IV. VOD (SD) Star Wars VII. VOD (HD) Inception CatchUp feb 17 Inception HBO2 22:00 Friends S01E01 SVOD Inception PVR 20:00 Inception HBO 20:00 News BBC HD 07:00 Items HBO SD HBO 2 SD Friends Series Movie Channels Products Inception (HD) Star Wars Package ($11.99) ($9.99) ($2.99) Inception TVOD (19:25; 1136s) (19:00) (21:11; 3651s; 5*) (16:13; 3715s) (20:10; 1135s) Interface TV provider related space BBC HD
  • 19. Friends S01E01 Friends S02E05 Consumers Friends S01E02 News 2016-02- 18 07:00 (Mary) (John) (Mark) Contents Star Wars IV. VOD (SD) Star Wars VII. VOD (HD) Inception CatchUp feb 17 Inception HBO2 22:00 Friends S01E01 SVOD Inception PVR 20:00 Inception HBO 20:00 News BBC HD 07:00 Items HBO SD HBO 2 SD Friends Series Movie Channels Products Inception (HD) Star Wars Package ($11.99) ($9.99) ($2.99) Inception TVOD (19:25; 1136s) (19:00) (21:11; 3651s; 5*) (16:13; 3715s) (20:10; 1135s) Interface TV provider related space TopGear S01 Friends S01 Friends S02 Collections BBC HD
  • 20. Friends S01E01 Friends S02E05 Consumers Friends S01E02 News 2016-02- 18 07:00 (Mary) (John) (Mark) Contents Star Wars IV. VOD (SD) Star Wars VII. VOD (HD) Inception CatchUp feb 17 Inception HBO2 22:00 Friends S01E01 SVOD Inception PVR 20:00 Inception HBO 20:00 News BBC HD 07:00 Items HBO SD HBO 2 SD Friends Series Movie Channels Products Inception (HD) Star Wars Package ($11.99) ($9.99) ($2.99) Inception TVOD (19:25; 1136s) (19:00) (21:11; 3651s; 5*) (16:13; 3715s) (20:10; 1135s) Interface TV provider related space TopGear S01 Friends S01 Friends S02 Collections Movies Series Shows Domains Domain dependent contents BBC HD
  • 21. Friends S01E01 Friends S02E05 Consumers Friends S01E02 News 2016-02- 18 07:00 (Mary) (John) (Mark) Contents Star Wars IV. VOD (SD) Star Wars VII. VOD (HD) Inception CatchUp feb 17 Inception HBO2 22:00 Friends S01E01 SVOD Inception PVR 20:00 Inception HBO 20:00 News BBC HD 07:00 Items HBO SD HBO 2 SD Friends Series Movie Channels Products Inception (HD) Star Wars Package ($11.99) ($9.99) ($2.99) Inception TVOD (19:25; 1136s) (19:00) (21:11; 3651s; 5*) (16:13; 3715s) (20:10; 1135s) Interface TV provider related space TopGear S01 Friends S01 Friends S02 Collections Movies Series Shows Domains TV Domain dependent contents Musics Books BBC HD
  • 22. Movies Series Shows Friends S01E01 Friends S02E05 Star Wars IV. VOD (SD) Inception (HD) Star Wars VII. VOD (HD) Inception CatchUp feb 17 Inception HBO2 22:00 Friends S01E01 SVOD Inception PVR 20:00 Male Action Inception HBO 20:00 Domain dependent contents Collections Contents Items Products Interface Consumers TV Friends S01E02 Domains TV provider related space News 2016-02- 18 07:00 News News BBC HD 07:00 Harrison Ford Female Comedy HBO SD HBO 2 SD HD SD Friends Series Star Wars Package TopGear S01 Friends S01 Friends S02 Musics (19:25; 1136s) (19:00) (16:13; 3715s) ($11.99) Movie Channels ($9.99) (8880s; 20:00-22:28) (Mary) (John) (Mark) (20:10; 1135s) Inception TVOD ($2.99) (21:11; 3651s; 5*) BBC HD Meta data Books
  • 23. Movies Series Shows Friends S01E01 Friends S02E05 Star Wars IV. VOD (SD) Inception (HD) Star Wars VII. VOD (HD) Inception CatchUp feb 17 Inception HBO2 22:00 Friends S01E01 SVOD Inception PVR 20:00 Male Action Inception HBO 20:00 Domain dependent contents Collections Contents Items Products Interface Consumers TV Friends S01E02 Domains TV provider related space News 2016-02- 18 07:00 News News BBC HD 07:00 Harrison Ford Female Comedy HBO SD HBO 2 SD HD SD Friends Series Star Wars Package TopGear S01 Friends S01 Friends S02 Musics (19:25; 1136s) (19:00) (16:13; 3715s) ($11.99) Movie Channels ($9.99) (8880s; 20:00-22:28) (Mary) (John) (Mark) (20:10; 1135s) Inception TVOD ($2.99) (21:11; 3651s; 5*) BBC HD Meta data Books
  • 24. 24 Graph representation Nodes • Entities: Users, items, contents, product, collections • Metadata: Genres, actors, category tree nodes, gender, age group • Node properties: Auxiliary meta data (e.g. descriptions) Edges • Interactions: Implicit/explicit feedbacks • Metadata containment, relations • Edge properties: Context, direction
  • 27. 27 Evolution of the Recommender Problem Source: http://www.slideshare.net/xamat/recsys-2014-tutorial-the-recommender-problem-revisited
  • 28. 28
  • 30. Movies Series Shows Friends S01E01 Friends S02E05 Star Wars IV. VOD (SD) Inception (HD) Star Wars VII. VOD (HD) Inception CatchUp feb 17 Inception HBO2 22:00 Friends S01E01 SVOD Inception PVR 20:00 Male Action Inception HBO 20:00 Domain dependent contents Collections Contents Items Products Interface Consumers TV Friends S01E02 Domains TV provider related space News 2016-02- 18 07:00 News News BBC HD 07:00 Harrison Ford Female Comedy HBO SD HBO 2 SD HD SD Friends Series Star Wars Package TopGear S01 Friends S01 Friends S02 Musics (19:25; 1136s) (19:00) (16:13; 3715s) ($11.99) Movie Channels ($9.99) (8880s; 20:00-22:28) (Mary) (John) (Mark) (20:10; 1135s) Inception TVOD ($2.99) (21:11; 3651s; 5*) BBC HD Meta data Books
  • 31. Movies Series Shows Friends S01E01 Friends S02E05 Star Wars IV. VOD (SD) Inception (HD) Star Wars VII. VOD (HD) Inception CatchUp feb 17 Inception HBO2 22:00 Friends S01E01 SVOD Inception PVR 20:00 Male Action Inception HBO 20:00 Domain dependent contents Collections Contents Items Products Interface Consumers TV Friends S01E02 Domains TV provider related space News 2016-02- 18 07:00 News News BBC HD 07:00 Harrison Ford Female Comedy HBO SD HBO 2 SD HD SD Friends Series Star Wars Package TopGear S01 Friends S01 Friends S02 Musics (19:25; 1136s) (19:00) (16:13; 3715s) ($11.99) Movie Channels ($9.99) (8880s; 20:00-22:28) (Mary) (John) (Mark) (20:10; 1135s) Inception TVOD ($2.99) (21:11; 3651s; 5*) BBC HD Meta data Books
  • 32. Star Wars IV. VOD (SD) Inception (HD) Star Wars VII. VOD (HD) Inception CatchUp feb 17 Inception HBO2 22:00 Friends S01E01 SVOD Inception PVR 20:00 Inception HBO 20:00 Items Products Interface Consumers News BBC HD 07:00 HBO SD HBO 2 SD Friends Series Star Wars Package (19:25; 1136s) (19:00) (16:13; 3715s) ($11.99) Movie Channels ($9.99) (8880s; 20:00-22:28) (Mary) (John) (Mark) (20:10; 1135s) Inception TVOD ($2.99) (21:11; 3651s; 5*) BBC HD
  • 33. Star Wars IV. VOD (SD) Inception (HD) Star Wars VII. VOD (HD) Inception CatchUp feb 17 Inception HBO2 22:00 Friends S01E01 SVOD Inception PVR 20:00 Inception HBO 20:00 Items Consumers News BBC HD 07:00 (8880s; 20:00-22:28) (Mary) (John) (Mark)
  • 34. 34 Collaborative Filtering / Graph representation The Matrix The Matrix 2 The Matrix 3 Twilight
  • 35. 35 Collaborative Filtering / Matrix representation The Matrix The Matrix 2 The Matrix 3 Twilight R The Matrix The Matrix 2 Twilight The Matrix 3 User 1 1 1 ? ? User 2 1 1 0 0 User 3 1 0 1 1 User 4 1 0 0 1 ? ?
  • 36. 36 Collaborative Filtering / Binary preference matrix R The Matrix The Matrix 2 Twilight The Matrix 3 User 1 1 1 ? ? User 2 1 1 0 0 User 3 1 0 1 1 User 4 1 0 0 1
  • 37. 37 Collaborative Filtering / Implicit MF R The Matrix The Matrix 2 Twilight The Matrix 3 User 1 User 2 𝒓 𝑢𝑖 User 3 User 4 Q 𝒒𝑖 PT 𝒑 𝑢 𝑇 𝑹 𝑵𝒙𝑴: preference matrix 𝑷 𝑵𝒙𝑲: user feature matrix 𝑸 𝑴𝒙𝑲: item feature matrix 𝑵: #users 𝑴: #items 𝑲: #features 𝑲 ≪ 𝑴 , 𝑲 ≪ 𝑵 𝑹~𝑷 𝐓 𝑸 𝑟 𝑢𝑖 = 𝒑 𝑢 T 𝒒𝑖
  • 38. 38 Collaborative Filtering / Implicit MF / Weighted Sum of Squared Errors R The Matrix The Matrix 2 Twilight The Matrix 3 User 1 1 1 ? ? User 2 1 1 0 0 User 3 1 0 1 1 User 4 1 0 0 1 C The Matrix The Matrix 2 Twilight The Matrix 3 User 1 c11 c12 1 1 User 2 c21 c22 1 1 User 3 c31 1 c33 c34 User 4 c41 1 1 1 𝒇 𝑷, 𝑸 = 𝑾𝑺𝑺𝑬 = (𝒖,𝒊) 𝒄 𝒖𝒊 𝒓 𝒖𝒊 − 𝒓 𝒖𝒊 𝟐• Objective Function:
  • 39. 39 Collaborative Filtering / Implicit MF / Alternating Least Squares R The Matrix The Matrix 2 Twilight The Matrix 3 User 1 User 2 User 3 User 4 Q 0.1 -0.4 0.8 0.3 0.6 0.7 -0.1 0.8 PT -0.2 0.6 0.6 0.4 0.7 -0.2 0.5 -0.2 𝑹~𝑷 𝐓 𝑸 𝑟 𝑢𝑖 = 𝒑 𝑢 T 𝒒𝑖 Ridge Regression • 𝑝 𝑢 = 𝑄𝐶 𝑢 𝑄T −1 𝑄𝐶 𝑢 𝑅 𝑟 𝑢 • 𝑞𝑖 = 𝑃𝐶 𝑖 𝑃T −1 𝑃𝐶 𝑖 𝑅 𝑐 𝑖
  • 40. 40 Collaborative Filtering / Implicit MF / Alternating Least Squares R The Matrix The Matrix 2 Twilight The Matrix 3 User 1 User 2 User 3 User 4 Q 0.3 -0.3 0.6 0.2 0.7 0.8 0.2 0.7 PT -0.2 0.6 0.6 0.4 0.7 -0.2 0.5 -0.2 Ridge Regression • 𝑝 𝑢 = 𝑄𝐶 𝑢 𝑄T −1 𝑄𝐶 𝑢 𝑅 𝑟 𝑢 • 𝑞𝑖 = 𝑃𝐶 𝑖 𝑃T −1 𝑃𝐶 𝑖 𝑅 𝑐 𝑖 𝑹~𝑷 𝐓 𝑸 𝑟 𝑢𝑖 = 𝒑 𝑢 T 𝒒𝑖
  • 41. 41 Collaborative Filtering / Implicit MF / Alternating Least Squares R The Matrix The Matrix 2 Twilight The Matrix 3 User 1 User 2 User 3 User 4 Q 0.3 -0.3 0.6 0.2 0.7 0.8 0.2 0.7 PT -0.2 0.7 0.6 0.5 0.8 -0.2 0.6 -0.2 Ridge Regression • 𝑝 𝑢 = 𝑄𝐶 𝑢 𝑄T −1 𝑄𝐶 𝑢 𝑅 𝑟 𝑢 • 𝑞𝑖 = 𝑃𝐶 𝑖 𝑃T −1 𝑃𝐶 𝑖 𝑅 𝑐 𝑖 𝑹~𝑷 𝐓 𝑸 𝑟 𝑢𝑖 = 𝒑 𝑢 T 𝒒𝑖
  • 42. 42 Collaborative Filtering / Implicit MF / Alternating Least Squares R The Matrix The Matrix 2 Twilight The Matrix 3 User 1 0.02 0.45 User 2 User 3 User 4 Q 0.3 -0.3 0.6 0.2 0.7 0.8 0.2 0.7 PT -0.2 0.7 0.6 0.5 0.8 -0.2 0.6 -0.2 𝑹~𝑷 𝐓 𝑸 𝑟 𝑢𝑖 = 𝒑 𝑢 T 𝒒𝑖
  • 44. Movies Series Shows Friends S01E01 Friends S02E05 Star Wars IV. VOD (SD) Inception (HD) Star Wars VII. VOD (HD) Inception CatchUp feb 17 Inception HBO2 22:00 Friends S01E01 SVOD Inception PVR 20:00 Male Action Inception HBO 20:00 Domain dependent contents Collections Contents Items Products Interface Consumers TV Friends S01E02 Domains TV provider related space News 2016-02- 18 07:00 News News BBC HD 07:00 Harrison Ford Female Comedy HBO SD HBO 2 SD HD SD Friends Series Star Wars Package TopGear S01 Friends S01 Friends S02 Musics (19:25; 1136s) (19:00) (16:13; 3715s) ($11.99) Movie Channels ($9.99) (8880s; 20:00-22:28) (Mary) (John) (Mark) (20:10; 1135s) Inception TVOD ($2.99) (21:11; 3651s; 5*) BBC HD Meta data Books
  • 45. Star Wars IV. VOD (SD) Inception (HD) Star Wars VII. VOD (HD) Inception CatchUp feb 17 Inception HBO2 22:00 Friends S01E01 SVOD Inception PVR 20:00 Action Inception HBO 20:00 Contents Items Products Interface Consumers News News BBC HD 07:00 Harrison Ford Comedy HBO SD HBO 2 SD HD SD Friends Series Star Wars Package (16:13; 3715s) Movie Channels (8880s; 20:00-22:28) (John) (20:10; 1135s) Inception TVOD BBC HD
  • 46. Star Wars IV. VOD (SD) Inception (HD) Star Wars VII. VOD (HD) Inception CatchUp feb 17 Inception HBO2 22:00 Friends S01E01 SVOD Inception PVR 20:00 Action Inception HBO 20:00 Contents Items Products Consumers News News BBC HD 07:00 Harrison Ford Comedy HD SD (8880s; 20:00-22:28) (John)
  • 47. 47 Content-based Filtering / Cosine Similarity (1/6) Sci-Fi RomanceAdventure The Matrix The Matrix 2 The Matrix 3 Twilight
  • 48. 48 Content-based Filtering / Cosine Similarity (2/6) Sci-Fi RomanceAdventure The Matrix The Matrix 2 The Matrix 3 Twilight ? ? MT Sci-Fi Adventure Romance The Matrix 1 1 0 The Matrix 2 1 1 0 Twilight 0 1 1 The Matrix 3 1 1 0 R The Matrix The Matrix 2 Twilight The Matrix 3 User 1 1 ? ?
  • 49. 49 Content-based Filtering / Cosine Similarity (3/6) MT Sci-Fi Adventure Romance The Matrix 1 1 0 The Matrix 2 1 1 0 Twilight 0 1 1 The Matrix 3 1 1 0 R The Matrix The Matrix 2 Twilight The Matrix 3 User 1 1 ? ?
  • 50. 50 Content-based Filtering / Cosine Similarity (4/6) M’T Sci-Fi Adventure Romance The Matrix 1/6 1/8 0 The Matrix 2 1/6 1/8 0 Twilight 0 1/8 1/2 The Matrix 3 1/6 1/8 0 R The Matrix The Matrix 2 Twilight The Matrix 3 User 1 1 ? ? 𝑚′𝑖𝑗 = 𝑚𝑖𝑗 𝑘 𝑚𝑖𝑘 𝑘 𝑚 𝑘𝑗
  • 51. 51 Content-based Filtering / Cosine Similarity (5/6) M’T Sci-Fi Adventure Romance The Matrix 1/6 1/8 0 The Matrix 2 1/6 1/8 0 Twilight 0 1/8 1/2 The Matrix 3 1/6 1/8 0 R The Matrix The Matrix 2 Twilight The Matrix 3 User 1 1 ? ? PT Sci-Fi Adventure Romance User 2/6 2/8 0 𝑃T = 𝑅𝑀′T
  • 52. 52 Content-based Filtering / Cosine Similarity (6/6) M’T Sci-Fi Adventure Romance The Matrix 1/6 1/8 0 The Matrix 2 1/6 1/8 0 Twilight 0 1/8 1/2 The Matrix 3 1/6 1/8 0 R The Matrix The Matrix 2 Twilight The Matrix 3 User 1 1 ? ? PT Sci-Fi Adventure Romance User 2/6 2/8 0 𝑹 The Matrix The Matrix 2 Twilight The Matrix 3 User 1 1 0.268 1 𝑟𝑢𝑖 = 𝑘 𝑚′𝑖𝑘 𝑝 𝑢𝑘 𝑘 𝑚′𝑖𝑘 2 𝑘 𝑝 𝑢𝑘 2 𝑃T = 𝑅𝑀′T
  • 54. 54 • Who is the front of the TV? (watching behavior pattern recognition) • Cold-start problem, opt-out users (metadata enrichment, hybrid filtering) • Noise filtering, lack of explicit data (filtering and weighting method, implicit feedbacks) • Context-dependency (device, time, mood) • Cross-domain recommendation, content centralization • Time-dependent recommendable set (model-based methods) • Popularity effect (diversification) • Upselling, subscription (Live  VOD) • Offline vs. Online KPIs Data Science Challenges
  • 55. 55 • Heterogeneous external data sources (data unification) • Big data (scalable algorithms, distributed databases) • Latency in data transfer, streaming service • Responsivity, load, SLA (distributed systems, load balancer) • Maintenance, service, follow-up • Open source vs. self-development • Distributed systems vs. single server • Software-as-a-service vs. on-site deployment Technology Challenges
  • 56. 56 Presented by: Contact: THANK YOU! www.impresstv.com David Zibriczky Head of Data Science david.zibriczky@impresstv.com
  • 57. 57 • Senior Software Engineers (Java, .NET) • Developer (Java, Python, Node.js) • Senior Data Mining Engineer • Pre-sales Technical Consultant • Data Scientist • Contact: hr@impresstv.com • Website: www.impresstv.com We are hiring!