g:i ag:i aAIM Analytics: SungJin Namg:i a - g:i a


Event Details

Predicting Disengaged Behaviors in an Online Meaning-Generation Task

Please RSVP Here

In an intelligent tutoring system (ITS), knowing when a student has disengaged from a task can be important for identifying the right moment of intervention. However, real-time prediction of disengagement is a challenging problem in complex cognitive domains like vocabulary learning. SungJin Nam will discuss data-driven methods to address two aspects of this problem: identification of predictive features and selection of accurate and robust models.

In this study, experiment data was collected in a middle-school classroom using a vocabulary training ITS. On each trial, the ITS presented a low-frequency (novel or frontier) word, and prompted students to generate the word’s meaning. Using log data collected in the ITS, this study showed quantitative interaction measures and linguistic characteristics of student responses can be used to predict the student’s potential disengaged behaviors while using the system. First, as expected, different types of features derived from log data effectively predicted students’ disengaged behaviors. Second, features based on relationship between responses were helpful for predicting different types of disengaged responses, such as repetitive or sequentially constructed responses. These findings provide useful insights on the representation and modeling of real-time changes in student engagement in ITS. Such models may be useful in the design and implementation of adaptive tutors in complex domains like language learning.

SungJin NamSungJin Nam received a bachelor of arts degree in psychology and information and culture technology from Seoul National University and a master of science degree in information from the University of Michigan. He is currently a doctoral candidate in the School of Information at the University of Michigan. With a background in psychology and data science, he is interested in interpreting human behavior using data-driven methods.

Please RSVP Here


() g:i a - g:i a

Privacy Preference Center