multi-source provenance-aware user interest profiling on the social semantic web
DESCRIPTION
UMAP 2012 Doctoral ConsortiumTRANSCRIPT
Copyright 2011 Digital Enterprise Research Institute. All rights reserved.
Digital Enterprise Research Institute www.deri.ie
Enabling Networked Knowledge
Multi-Source Provenance-Aware
User Interest Profiling
on the Social Semantic Web
Fabrizio Orlandi
Doctoral Consortium – UMAP 2012
Digital Enterprise Research Institute www.deri.ie
Enabling Networked Knowledge
Research Goal
Improve the current user interest profiling
techniques leveraging:
Linked Data,
Provenance of Data,
the Social Semantic Web.
2
Digital Enterprise Research Institute www.deri.ie
Enabling Networked Knowledge
The Web of Data
Linking Open Data cloud diagram, by Richard Cyganiak and Anja Jentzsch.
Digital Enterprise Research Institute www.deri.ie
Enabling Networked Knowledge
The Web of Data
db:Montreal
db:Quebec
db:Gilles_Villeneuve
db:Ferrari db:Formula_1
dbo:wikiPageWikiLinkdbo:wikiPageWikiLink
dbo:birthPlace
dbp:largestcity
Digital Enterprise Research Institute www.deri.ie
Enabling Networked Knowledge
Research Areas
Social media integration and interoperability
How to extract and aggregate relevant user information from social media
websites and make it available following the Linked Data principles?
How adaptive should be a user profiling algorithm according to the type of social
media?
Provenance of data
What is the role of provenance on the Social Web and on the Web of Data and how
to use it for user profiling?
How dependent are profiling algorithms from the origin, history and types of user
activities on Social Web and how to adapt to it?
The Web of Data for interest profiling
How to use the Web of Data and semantic technologies to enrich user profiles?
How to leverage the Web of Data for different ranking strategies of user interests?
5
Digital Enterprise Research Institute www.deri.ie
Enabling Networked Knowledge
Challenges – 1
Information on the Social Web is stored in isolated data silos
on heterogeneous and disconnected social media websites
http://www.w3.org6
Digital Enterprise Research Institute www.deri.ie
Enabling Networked Knowledge
Challenges – 1
User profiles should be represented in an interoperable way
in order to exchange information across different systems
[image: U. Bojārs, A. Passant, J. Breslin]7
Digital Enterprise Research Institute www.deri.ie
Enabling Networked Knowledge
Research Questions
Social media integration and interoperability
How to extract and aggregate relevant user information from social
media websites and make it available following the Linked Data
principles?
How adaptive should be a user profiling algorithm according to the type
of social media?
Provenance of data
What is the role of provenance on the Social Web and on the Web of Data and how to use it for
user profiling?
How dependent are profiling algorithms from the origin, history and types of user activities on
Social Web and how to adapt to it?
The Web of Data for interest profiling
How to use the Web of Data and semantic technologies to enrich user profiles?
How to leverage the Web of Data for different ranking strategies of user interests?
8
Digital Enterprise Research Institute www.deri.ie
Enabling Networked Knowledge
Challenges – 2
Lack of provenance on the Web of Data:
datasets on the Social Web are often the result of data
mashups or collaborative user activities
9
Digital Enterprise Research Institute www.deri.ie
Enabling Networked Knowledge
Research Questions
Social media integration and interoperability
How to extract and aggregate relevant user information from social media websites and make it
available following the Linked Data principles?
How adaptive should be a user profiling algorithm according to the type of social media?
Provenance of data
What is the role of provenance on the Social Web and on the Web of Data
and how to use it for user profiling?
How dependent are profiling algorithms from the origin, history and
types of user activities on Social Web and how to adapt to it?
The Web of Data for interest profiling
How to use the Web of Data and semantic technologies to enrich user profiles?
How to leverage the Web of Data for different ranking strategies of user interests?
10
Digital Enterprise Research Institute www.deri.ie
Enabling Networked Knowledge
Challenges – 3
The Web of Data: a continuously evolving “open corpus”
LOD Cloud by R. Cyganiak
and A. Jentzsch11
Digital Enterprise Research Institute www.deri.ie
Enabling Networked Knowledge
Research Questions
Social media integration and interoperability
How to extract and aggregate relevant user information from social media websites and make it
available following the Linked Data principles?
How adaptive should be a user profiling algorithm according to the type of social media?
Provenance of data
What is the role of provenance on the Social Web and on the Web of Data and how to use it for
user profiling?
How dependent are profiling algorithms from the origin, history and types of user activities on
Social Web and how to adapt to it?
The Web of Data for interest profiling
How to use the Web of Data and semantic technologies to enrich user
profiles?
How to leverage the Web of Data for different ranking strategies of
user interests?
12
Digital Enterprise Research Institute www.deri.ie
Enabling Networked Knowledge
Outline
The user profiling data process:
1. from user activities on heterogeneous social media websites,
2. to their provenance representation,
3. to the data aggregation, analysis and integration with the Web of Data.
12
3
13
Digital Enterprise Research Institute www.deri.ie
Enabling Networked Knowledge
Work done
Aggregated, Interoperable and Multi-Domain
User Profiles of Interests for the Social Web
Privacy Aware and Faceted
User-Profile Management
Personalized Filtering of
the Twitter Stream
14
Semantic integration of social networking
platforms (the wikis use case)
Semantic representation and management of provenance on the
Social Web and the Web of Data (DBpedia)
Month:
1st – 6th
6th – 18th
18th – 24th
Digital Enterprise Research Institute www.deri.ie
Enabling Networked Knowledge
Aggregated, Interoperable and Multi-
Domain User Profiles for the Social Web
15
Digital Enterprise Research Institute www.deri.ie
Enabling Networked Knowledge
16
Open Questions
How adaptive should be a user profiling algorithm according to the
type of social media?
What are the differences between extracting user interests on Microblogs, Wikis,
Social Networking sites, etc.?
How can a general purpose user interesting profiling algorithm adapt to it?
How dependent are profiling algorithms from the origin, history and
types of user activities on Social Web and how to adapt to it?
What are the different types of activities that users perform on the Social Web
expressing personal interest and how to weight them?
How does detailed provenance information about user activities help in creating
more accurate and fine-grained profiles?
How to leverage the Web of Data for different ranking strategies of
user interests?
How relevant are the collected interests for a user profile and what are their
relations with other concepts on the Web of Data?
Digital Enterprise Research Institute www.deri.ie
Enabling Networked Knowledge
Future Work
■ User profiling on Wikipedia analysing authorship and contributions
for DBpedia statements and Wikipedia articles.
■ Test of user interest profiling strategies on different scenarios
(Microblogs, Wikis, etc.)
■ Integration and enrichment of the semantic user profiles generated
with the Web of Data and other Social Media
■ Evaluation of the generated user profiles
17
Digital Enterprise Research Institute www.deri.ie
Enabling Networked Knowledge
Thanks
Contacts:
http://bit.ly/M7hvbX
@BadmotorF
18
Thanks to:
Alexandre Passant - @terraces
John Breslin - @johnbreslin