Affective modelling and feedback in programming practice systems

College

College of Computer Studies

Document Type

Dissertation

Publication Date

3-2020

Abstract

Affective modelling and feedback have been shown to be potentially useful in intelligent tutoring systems. This is based on several studies showing that emotions experienced by students are correlated with various aspects of learning. The computer science community have explored ways to model and respond to student emotions in several learning domains. In my dissertation work, I focus on modelling and responding to the emotional states of university students while doing coding exercises. In this type of activity, the student acts as an individual programmer writing code alone, a setup like when a student is doing practice at home without teacher supervision. In this kind of setup, the display of emotions is more challenging to detect than that of more traditional tutoring interactions because they are more subtle and naturalistic. To address this, I use a combination of face features and system log features to train models to estimate emotion while coding. I then use these models to investigate simple affective feedback in systems for programming practice, such as generating problems and offering guides based on confusion, as well as providing emotional responses based on the affective state of the student. We found that some log features and some face features are associated with certain emotional states in programming and can be combined to train models with a slight improvement over previous approaches. We also developed two systems with simple affective feedback, EmoTutor1 and EmoTutor2, and found that these systems can help students solve more problems and have a more positive impression in terms of learning experience and engagement, when compared to traditional methods that do not provide feedback. However, the timely presentation of such simple interventions was not found to be a significant factor in those positive effects.

html

Disciplines

Computer Sciences

Keywords

Intelligent tutoring systems

Upload File

wf_no

This document is currently not available here.

Share

COinS