17
Personalized News Recommendation System with Different User Modeling Engineers: Mahesh Attarde (201205685) Arpita Raj Gupta (201101121) Pankhuri Goyal (201101174)

Personalized News Recommendation based on Twitter User Modeling

Embed Size (px)

Citation preview

Page 1: Personalized  News Recommendation based on Twitter User Modeling

Personalized News

Recommendation System with Different

User Modeling

Engineers: Mahesh Attarde (201205685) Arpita Raj Gupta (201101121) Pankhuri Goyal (201101174)

Page 2: Personalized  News Recommendation based on Twitter User Modeling

Problem

INFORMATION OVERLOAD

Understand “Me”

Recommend news that “Interest”

Me!

Use Social My “Social

Behavior”

Learn Me over “Time”

Page 3: Personalized  News Recommendation based on Twitter User Modeling

Problem

INFORMATION OVERLOAD

Understand “Me”

Recommend news that “Interest”

Me!

Use Social My “Social

Behavior”

Learn Me over “Time”

Personalized News Recommendation

Page 4: Personalized  News Recommendation based on Twitter User Modeling

Social Feed -> Web

Tweet Processing

Enrichment

User Analysis User Models

URL enrichment

Topic/Entity Enrichment

Entity/HashTag/Topic Identification

Temporal Classification

Personalized

Page 5: Personalized  News Recommendation based on Twitter User Modeling

Challenge 1

• Design of User Model

[across million attributes and users]

• Design Similarity of Users

User Model

Page 6: Personalized  News Recommendation based on Twitter User Modeling

Challenge 2

• Design of Recommendation

[filtering across million attributes and users]

• Design ranking in user attributes [topic, entity, hash tag ]

Recommendation

Page 7: Personalized  News Recommendation based on Twitter User Modeling

User Model -> Rec Engine

• Temporal Feature

• Profile Types

• Enrichment Types

Page 8: Personalized  News Recommendation based on Twitter User Modeling

Prototype – What we had

• Twitter 7.6 GB corpus

• Over 1.2 years

• 1218 users with avg.

of 175 tweets per user

• News corpus of 75,000+ articles

Page 9: Personalized  News Recommendation based on Twitter User Modeling

Prototype – What we Found

• Hash Tags distinct 100000+

• Topics - distinct 18 types

• Entities – distinct 39 types

Temporal Feature of Hash Tag

Page 10: Personalized  News Recommendation based on Twitter User Modeling

Prototype – our Play

• User Profiles

1)Hash Tag Profile

2)Entity Profile

3)Topic Profile

4)Experimental enriched

• Recommendation Engine cosine similarity based

Page 11: Personalized  News Recommendation based on Twitter User Modeling

Results

• Hash tag profiles grow quickly and last for period [Used Weka-Excel]

• Entity and Topic Profiles suggest relevant

suggestions

• Enrichment in Topics and Entity help to correctly judge “subject”

<Improved Profile>

TEMPORAL EFFECT

PROFILE EFFECT

Enrichment EFFECT

Page 12: Personalized  News Recommendation based on Twitter User Modeling

Snap Shots

Page 13: Personalized  News Recommendation based on Twitter User Modeling

Snap Shots

Page 14: Personalized  News Recommendation based on Twitter User Modeling

Snap Shots

Page 15: Personalized  News Recommendation based on Twitter User Modeling

Snap Shots

Page 16: Personalized  News Recommendation based on Twitter User Modeling

Tools and Lib

• Weka

• Excel

• Open Cailais

• Stanford NPL lib

• Apache Mahout

• RDBMS -mysql

Page 17: Personalized  News Recommendation based on Twitter User Modeling

Thank You !

Professor Vasudev Varma

Aishvarya Singh

Guided By

Mentored By