Open User Models in Social Media

A user model is a representation of information about an individual user that is essential for personalization tasks. User models can be leveraged to generate effective recommendations and provide intelligent feedback. Further, opening these models to users and visualizing them in the right way could improve transparency, increase awareness, and promote self-reflection. The most popular features in user modeling include user’s knowledge, goals, background, individual traits, interests, and context. As access to information and participation in social media environments are mostly interest-driven and context-sensitive, user interests and contexts need to constitute important user features to be modeled in social media. This is a highly challenging task since user activities are often distributed across social media.  

This PhD project will investigate effective and efficient methods to aggregate, manage, analyze, visualize, and leverage the interest and context dimensions of user models in order to better induce and support personalization in social media (e.g. scientific communities, blogger networks, open source software developer groups, learning networks). Thereby, several issues have to be taken into account, including questions about privacy, scalability, integration, and interoperability.

Main objectives / research questions:

  • Develop new methods based on text mining, natural language processing (NLP), topic modeling, and semantic keyphrase extraction for an effective modeling of the user’s interests and contexts across different social media environments.
  • Design, implement, and evaluate a user-centric and privacy-preserving platform to collect, store, manage and query user model data.
  • Develop new methods based on machine learning (with human in the loop) and visual analytics to better explore the potential of the interest and context dimensions of user models to support awareness, reflection, feedback, intelligent search, and interactive recommendation in social media.

For more information, contact Prof. Dr. Mohamed Chatti