Research Interests

Currently, he focuses the research on user modeling, behavior analysis, human factors (user emotions and personalities), context-awareness, multi-criteria decision making, educational learning/learning analytics, and recommender systems. But he is also interested in the following topics:

  • HCI and Personalization: Recommender Systems, User Modeling, Human Factors, Learning Analytics
  • Data Science: AI/DM/DL, Business Intelligence, Decision Science, Learning Science


Below is a list of students or researchers under his supervision

  • List of PhD Students
    • Diego Sánchez-Moreno (Part-time PhD Student); Project: Music Recommender Systems
  • List of Master Students
    • Recommender Systems
        • Maria Delgado Franco, Mili Singh, Mayur Agnani; Project: Identification of Grey Sheep Users In Recommender Systems
        • Tanaya Dave, Neha Mishra, Harshit Kumar; Project: Reciprocal Recommendations
        • Shephalika Shekhar; Project: Multi-Criteria Recommender System
        • Alisha Anna Jose; Project: Context-Aware Recommender System
        • Nastaran Ghane, Milad Sabouri; Project: Multi-Stakeholder Recommendations
    • Cybersecurity
        • Raquel Noblejas Sampedro, Sridhar Srinivasan, Kim Taehun; Project: Malware Detection and Usage Analysis In Mobile Apps

Research Projects and Contributions

      • 2015 – Present, CARSKit: It is a Java-based open-source context-aware recommendation library.
      • 2014 – Present, Context Suggestion: The entity to be recommended is not longer a list of items, but a ranked list of appropriate context information, in order to maxmize user experiences.
      • 2014 – 2015, Similarity-Based Context Modeling: It provides a new way to incorporate context information into recommendation algorithms by learning similarity of contexts. And it is proved to work better than deviation-based context modeling approches (such as context-aware matrix factorization). See our papers in WISE 2015 and UMAP 2015
      • 2014, Contextual Sparse Linear Method (CSLIM): It contains a series of effective context-aware recommendation algorithms built upon Sparse Linear Method (SLIM) for top-N context-aware recommendations. See our papers in CIKM 2014 and RecSys 2014
      • 2012 – 2013, Differential Context Modeling (DCM): Differential context relaxation (DCR) and differential context weighting (DCW) are two hybrid context modeling approaches by using context information as filters. They are built upon neighborhood-based collaborative filtering (such as UserKNN, ItemKNN and slope one recommenders), and used to alleviate the sparsity problems in context-aware recommender systems.
      • 2011 – 2012, Experience Discovery: We work on the project which provides personalized recommendation services to information learning and after-school programs for all the public schools in the greate Chicago area. It is funded and supported by Chicago Public Schools, City of Learning and MacArthur Foundations.

Data Sets for Recommender Systems

Open Source Recommender Engines

    • Apache Mahout, Java-based library, especially for scalable data mining.
    • MyMediaLite, .Net-based library, specifically for recommender systems.
    • LensKit, Java-based recsys engine with classical recsys algorithms.
    • LibFM, a library for Factorization Machines
    • JavaFM, Java 8 Factorisation Machines Library
    • PITF, for tag recommendations
    • RankSys: Java 8 RecSys framework for novelty, diversity and much more
    • LibRec, Java-based library which implements the state-of-the-art recsys algorithms.
    • CARSKit, Java-based library for context-aware recommendations.