Learning Identifier: Unraveling the Attributes of Online Learning Behaviours

Learning Identifier: Unraveling the Attributes of Online Learning Behaviours


The age of information has drastically shifted the landscape of learning, with a significant leap towards online platforms. Internet browsing now encompasses a considerable fraction of self-driven, informal learning in addition to structured, formal online courses. In this context, the process of acquiring knowledge, understanding, behaviors, skills, values, attitudes, and preferences, otherwise known as learning, has become an intricate activity with dimensions extending beyond the traditional boundaries of the classroom. Despite its growing relevance, the characteristics of learning activities in an online environment, particularly in a browser setting, remain insufficiently mapped.

Our earlier work has set the foundation for this investigation. We have developed a Chrome browser extension called Tracker Toolbox, designed to collect and view user interaction statistics in real time. This tool enables data collection and visualization of various tracking metrics, such as eye movement patterns, activity duration, mouse interactions, and content tracking. The vast data generated through these metrics offer an opportunity to comprehend online learning behaviors in a granular and nuanced manner.



  • Behavior Modelling: One of the primary activities will be constructing behavior models that encapsulate online learning patterns. Using the data gathered by the Tracker Toolbox, we will construct detailed user behavior models. These models will reveal typical interaction patterns and sequences, which may signify distinct learning styles, engagement levels, and comprehension processes.
  • Search as Learning: A significant part of learning online involves searching for information. We will conduct a detailed analysis of users’ search behaviors, aiming to link specific search patterns to learning outcomes. This would involve studying search queries, sequence of queries, time spent on search results, and other relevant behaviors.
  • Machine Learning Classifiers: Leveraging machine learning, we will develop classifiers that can predict learning outcomes based on the observed user behaviors. These classifiers will be trained and validated using the data captured by the Tracker Toolbox. By using machine learning, we aim to create a system capable of identifying effective learning strategies and behaviors, as well as predicting possible challenges or obstacles to the learning process.


Project Team Members

  • Dr. Sharon Lynn Chu (ELX Lab Director)
  • Nanjie (Jimmy) Rao (Ph.D. Student, Computer Information Science & Engineering)
  • Ranger Chenore (Undergraduate Student, University of Florida)
  • Hang Ye (Master’s Student, University of Florida)