Recommending Insightful Actions In Learning Analytics Dashboards

The School of ITEE is hosting the following thesis review seminar:

Recommending Insightful Actions In Learning Analytics Dashboards

Speaker: Shiva Shabaninejad
Host: A/Prof. Hassan Khosravi

Abstract: The growing student population, demand for flexible course delivery, and increased use of educational technology have made it challenging for instructors to effectively realise and attend to students' needs. To help instructors make sense of student data, many learning analytics systems and dashboards have been proposed. At a high level, most of these systems either incorporate descriptive analytics where they rely on self-exploration and self-interpretation of the data from the instructors, which limits their functionality or they incorporate predictive analytics where the system makes automatic decisions without considering the instructor's judgment. To address these challenges, this thesis explores whether and how “augmented intelligence” can be employed in the context of learning analytics systems. Augmented intelligence is an approach that uses AI as a tool to augment and enhance human intelligence capabilities rather than operate independently. I present and evaluate three approaches based on augmented intelligence for enhancing existing descriptive and predictive learning analytics models.  The first method uses online analytical processing and decision trees to generate recommendations of insightful drill-downs to augment instructors with getting insights about sub-populations of students, based on the students learning characteristics. The second method incorporates process mining power to augment instructors with getting insights about sub-populations of students, based on students' learning processes. The third method uses explainable artificial intelligence techniques to rank students based on a set of risk factors to assist instructors with the identification and diagnostic analysis of at-risk students. I demonstrate the application of each method using real-life course data, and a prototype designed for an existing learning analytics dashboard. Furthermore, I report on a qualitative study with course instructors from a variety of disciplines that evaluates our methods' usefulness.

Speaker Biography: Shiva Shabaninejad is a Ph.D. candidate from the school of ITEE at the University of Queensland under the supervision of Assoc. Prof. Hassan Khosravi. Shiva holds a Master degree in Software and Systems from Polytechnic University of Madrid, and a bachelor degree in Software Engineering from Azad University of Tehran. Shiva has worked more than 7 years in industry as a software architect and developer and performed 3 years of research in academia before her Ph.D. study. Shiva’s lecturing expertise is in Database Management Systems, Managing Business Data, Data Warehousing and she has also experience in teaching Electronic Business Systems and Strategy as a casual lecturer at the Business School of the University of Queensland.  Shiva’s research interest include Learning Analytics, Educational Process Mining, Educational Data Mining, Process Mining and Machine Learning.