![User Profiling
“A user profile is a representation of information about an individual user
that is essential for the (intelligent) application we are considering” [1]
Contents of user profiles:
user interests;
the user’s knowledge, background and skills;
user behavior;
the user’s interaction preferences;
the user’s individual characteristics;
and the user’s context.
[1] S. Schiaffino, A. Amandi. 2009.](https://image.slidesharecdn.com/20120209knoesisfabrizio-120322111513-phpapp02/85/Semantic-user-profiling-and-Personalised-filtering-of-the-Twitter-stream-2-320.jpg)
![Challenges – 4
• User profiles should be represented in an interoperable way in order
to exchange information across different user adaptive systems
[U. Bojārs, A. Passant, J. Breslin]](https://image.slidesharecdn.com/20120209knoesisfabrizio-120322111513-phpapp02/85/Semantic-user-profiling-and-Personalised-filtering-of-the-Twitter-stream-7-320.jpg)
Semantic user profiling and Personalised filtering of the Twitter stream
- 1. User Profiling on the Social Semantic Web Fabrizio Orlandi, DERI (NUI Galway, Ireland) Kno.e.sis – WSU Dayton, OH – 9 Feb 2012
- 2. User Profiling “A user profile is a representation of information about an individual user that is essential for the (intelligent) application we are considering” [1] Contents of user profiles: user interests; the user’s knowledge, background and skills; user behavior; the user’s interaction preferences; the user’s individual characteristics; and the user’s context. [1] S. Schiaffino, A. Amandi. 2009.
- 3. Research Questions • How to collect and interlink user information from social media websites to build enhanced and comprehensive user profiles? • How to manage and merge user models from different applications and social sites in an interoperable way? • How to leverage provenance information and trust measures on the Web of Data to improve Web personalisation?
- 4. Challenges – 1 • Information on the Social Web is stored in isolated data silos on heterogeneous and disconnected social media websites http://www.w3.org
- 5. Challenges – 2 • The Web of Data: a continuously evolving “open corpus” LOD Cloud by R. Cyganiak and A. Jentzsch
- 6. Challenges – 3 • Lack of provenance on the Web of Data: datasets on the Social Web are often the result of data mashups or collaborative user activities
- 7. Challenges – 4 • User profiles should be represented in an interoperable way in order to exchange information across different user adaptive systems [U. Bojārs, A. Passant, J. Breslin]
- 8. Outline 1 3 2 The user profiling data process: 1. from user activities on heterogeneous social media websites, 2. to their provenance representation, 3. to the data aggregation and analysis
- 9. So far… State of the art analysis Modelling the structure of wikis Enabling semantic search on heterogeneous wiki systems Provenance of data in wikis Representation and extraction of provenance in Wikipedia and DBpedia Privacy Aware and Faceted User-Profile Management Personalized Filtering of the Twitter Stream…
- 10. Semantic Personalization of Social Web Streams
- 11. Motivation Twitter – Growth Information Overload 11 http://www.cmswire.com/cms/customer-experience/35-key-twitter-statistics-infographic-012384.php
- 15. Motivation • How many people should I follow? • Am I receiving latest/complete information? • How can I quickly tell the system what are my interests?
- 16. Approach -- Overview The new iPhone has a Broadcast 3.5-inch screen, released today Football User Profiles Filter Apple
- 17. Annotate: iPhone Get ?user foaf:interest Subscribers The new dbPedia:iPhone based on iPhone has a 3.5- Union preference inch screen, ?user foaf:interest released today Category:Apple Get Interested Subscribers Semantic Filter RDF Notify Update A N RDF N O Store and Query Topics Semantic T A T Fetch Updates Hub O RSS Store FOAF R Update RSS Profile Generator Push Updates to Interested Users Create Profile
- 18. User Profiling Interlink social websites Integration & Merge and model user data User Modelling User Profile Personalise users’ experience using their profile Recommendations Adaptive Systems Search Personalisation
- 19. User Profiling
- 20. Profile Generator • Data Extraction – Twitter, Facebook – Example: Tweets, FB Likes, posts, videos, etc. • Profile Generation – Interests extracted from collected data • Entity spotting (user generated data) • Explicit interests specified by user (Facebook likes etc.) – Weighted Interests w/ DBpedia resources/categories – FOAF profile
- 21. Semantic Filter Get Interested Subscribers RDF Semantic Filter Notify Update A N N O Store and Query Topics RDF Semantic T A T Fetch Updates Hub O RSS Store FOAF R Update RSS Profile Generator Create Profile
- 22. Semantic Filter • Twitter Storm: – Distributed realtime computation system • Microblog Metadata – Twitter provides metadata • Author, date, location etc.. – Metadata Extracted • DBpedia Entities, URLs • Generate SPARQL Query representing interested Users – Retrieved at Semantic Hub
- 23. Semantic Hub Get Interested Subscribers RDF Semantic Filter Notify Update A N N O Store and Query Topics RDF Semantic T A T Fetch Updates Hub O RSS Store FOAF R Update RSS Profile Generator Create Profile
- 24. Semantic Hub • RSS Extension – Preference – to include the SPARQL queries • Push content – FOAF profiles of the subscribers are matched with the preference – Interested subscribers receive the content
- 25. DERI’s Unit for Social Software (USS) Unit leader: John Breslin
- 26. Overview of research activities • Research team at DERI – Two postdocs (plus one starting on Monday) • Alex Passant (10%), Maciej Dabrowski, Bahareh Heravi – Nine PhD students • Six supervised by John, two by Alex, one by Michael H • Various interdisciplinary collaborations – Exercise, e-government, political science, journalism
- 27. Current students David Crowley Ted Vickey • Citizen sensors • Exercise adherence via – Funded by College of social networks Engineering and Informatics – Funded by American Council • Attaching data from on Exercise and IRCSET sensors to social web • Developing a classification content using semantic for fitness tweets to see if technologies sharing exercise regimes can encourage others
- 28. Current students Antonio Aguilar (EEE) Fabrizio Orlandi • Heart rate variability • User profiling on the Social analysis Semantic Web – Funded by Assisted Ambient – Funded by Cisco Foundation Living eCAALYX EU project and IRCSET • Developing methods to • Consolidating user profiles help predict sudden from various platforms and cardiac death using non- deriving interests from linear algorithms amalgamation
- 29. Current students Lukasz Porwol Owen Sacco • e-Participation via social • Trust, accountability and media privacy via Linked Data – Funded by Science – Funded by Cisco Foundation Foundation Ireland and IRCSET • Leveraging popular • Developing privacy networks for e-government preference managers for instead of standalone the Semantic Web platforms • Collaboration with US Government
- 30. Current students Marie Boran Jodi Schneider • Connecting data journalists • Argumentative discussions with linked scientific data – Funded by Science – Funded by Science Foundation Ireland Foundation Ireland • Representing, classifying • Bridging the gap between and visualizing experimental data from argumentative discussions scientists and the on the Web mainstream media
- 31. Current students Myriam Leggieri • Linked sensor data – Funded by SPITFIRE • Connecting sensor data with explanatory facts from the Linked Open Data Cloud
- 32. Some past postgraduate students • Sheila Kinsella – ECE graduate, now engineer with Datahug • Haklae Kim – Now senior engineer with Samsung • Uldis Bojars – Now with the National Library of Latvia • Gerard Cahill – BSc IT graduate, now developer with Starlight
- 33. DERI – House DERI Applied Commercialisation Research eBusiness eLearning Financial Services Health Care Green & eGovernment Life Sciences Sustainable IT Linked Data Research Stream 1: Stream 2: Semantic Stream 3: Semantic Stream 4: Semantic Centre Semantic Search Collaboration Information Mining Middleware Information Sensor Reasoning and Semantic Colla- Mining Middleware Querying borative Software and Retrieval Data Intensive Natural Language Service Oriented Social Software Processing Architecture Infrastructure
- 34. Thanks! • Contacts: - [email protected] - Twitter: BadmotorF
- 36. Some additional stats… • On average for: – 200 Tweets – 200 Facebook posts, and items. • ~106 interests - DBpedia instances • ~720 interests - DBpedia categories (~6.8 times more) • Estimated average Recall: 0.74 • 22 users