Computer Science (COMP) 649

Affective Computing (Revision 1)

COMP 649 Course website

Permanently closed, effective April 25, 2018.

Delivery Mode: Individualized Study Online

Credits: 3

Area of Study: IS Elective

Prerequisite: COMP 501 or COMP 504 or approval from the course coordinator before registering.

Faculty: Faculty of Science and Technology

Centre: School of Computing and Information Systems

Instructor: Dr. Maiga Chang

**Note: This is a graduate level course and students need to apply and be approved to one of the graduate programs or as a non-program School of Computing and Information graduate student in order to take this course. Minimum Admission Requirements must be met. Undergraduate students who do not meet admission requirement will not normally be permitted to take this course.


Affective computing is not only an important research area that uses classifiers and text analysis to detect emotional states; it also supports research in many other areas such as human–computer interaction, adaptive and personalized systems, and educational technologies. By applying affective computing to information systems and agents, computers can become capable of detecting users’ emotional states and giving them the most appropriate services and responses. This is intended to make users feel comfortable interacting with computers as well as to improve their mood make and it quicker and easier to consume information/knowledge.

This course shows students how affective computing works, methods of detecting a user’s emotional states, how affective computing can build a computer’s emotional states, and how computers interact with users based on emotional state detection and changes. Students will research, design, integrate, refine, and implement the latest affective computing methods. They will also become familiar with technologies/topics of their choice, and dig deeper through explaining and discussing these with peers.

Course Objectives

  • Inform graduate students in MSc Information Systems of the background and directions of affective computing research
  • Focus on the text-based and action log analysis-based approaches.
  • Provide students with both broad and in-depth knowledge, and a critical understanding of affective computing from different viewpoints: sending affective information, interpreting affective information, and simulating internal emotions.
  • Give students the opportunity to design their own affect analysis model and method.

Learning Outcomes

After successfully completing this course, students should be able to

  • summarize the theories and principles of the latest affective computing models and methods;
  • explain how emotional states change over time;
  • compare the pros and cons of different text-based affect analysis methods;
  • design a text-based affect analysis model and method;
  • evaluate the effectiveness of the proposed model and method;
  • independently present and explain the proposed text-based affect analysis method with social software and an asynchronous/synchronous online meeting service;
  • effectively communicate course work in writing and oral presentation.


Unit 1: Getting Familiar with Social Software and Academic Search Services
Students will become familiar with the social software used later in this course, e.g., wikis, online meeting services, YouTube, and vShare. Students will be asked to locate academic literature via e-journals and database search functions using the AU Library website, Google Scholar, and Google.

Unit 2: Fundamentals of Affective Computing
Students will examine several important overviews of the fundamental ideas and research in affective computing. The unit also asks students to describe what they have learned and discuss it with each other in a course discussion forum. Instructors will point out the key researchers and organizations to help students find research directions and/or topics of interest to them.

Unit 3: Receiving, Interpreting, and Sending Affective Information
Students will examine the literature on the theories and methods of interpreting affective information from text. This unit also asks students to do an in-depth survey on the affect analysis methods they are interested in, summarize the foundational theory as well as the history of improving and enhancing specific methods, and write a report. This will help prepare students to design their own affect analysis methods in Unit 6.

Unit 4: Synthesizing and Simulating Internal Emotions
Students will examine the literature on emotion states and models. This unit asks students to design a model using the text-based affect analysis method to store and represent emotion states of software agents as well as users. At the end of this unit, they will do a presentation on their model design and corresponding affect analysis method. Students will provide feedback on peers’ presentations.

Unit 5: Social, Ethical, and Philosophical Issues
Students will examine the literature on the social and ethical issues around affective computing applications. Students will describe what they have learned and discuss it with each other in a course discussion forum.

Unit 6: Designing a Text-based Affect Analysis Model and Method
First, each student will complete a design of an emotion model and affect analysis method in response to the feedback from peers when they presented their idea in Unit 4. After that, they will implement the proposed model and method in Java (or in any programming language that can run on a web browser using Linux) to develop software agents that can talk with one or more users at the same time. They will evaluate their proposed model and method by talking to the agent online and observing the agent’s responses. They are also encouraged to discuss their observations and findings in a course discussion forum.

Unit 7: Presentation and Reflection
Students will demonstrate the emotion model and affect analysis method they designed and implemented starting in Unit 4. They will prepare presentation slides to use with an online meeting service that allows other students to join synchronously. The presentation will be recorded to allow it to be viewed asynchronously also. Students will answer any questions from peers during their presentation as well as any questions posted in a course discussion forum. Students will comment on the presentations that others have done and ask questions asynchronously or synchronously.


In order to receive credit for COMP 649, you must achieve a cumulative course grade of "C+" (66 percent) or better, and with an average grade of at least 66% for the Reading, Design, and Presentation assignments and an average grade of at least 66% for the Implementation and Demonstration assignments. Your cumulative course grade will be based on the following assessment.

Assessment Weight  
Assignment 1: Reading Report 10% Average grade should be at least 66%
Assignment 2: Design of Affect Analysis Model and Method 15%
Assignment 3: Design Presentation 10%
Assignment 4: Implementation of Affect Analysis Model and Method 30% Average grade should be at least 66%
Assignment 5: Final Presentation 20%
Participation 15%

Course Materials

Readings for this course will be taken entirely from web-based resources, which typically can be accessed via academic journal databases on the AU Library website, Google Scholar, and Google. Content will be discovered and accessed by the students. The most recent significant publications that have generated strong interest will be selected and added to the recommend reading materials, wiki, and course news feed.

Other Materials

The remainder of the learning materials is distrubted in the online learning environment.

Special Course Features

COMP 649 is offered online and can be completed at the student’s workplace or home. Students will use an asynchronous/synchronous online meeting service to present and discuss projects. They will need a webcam, and a microphone and speakers. A plug-in headset with a microphone is ideal.

Athabasca University reserves the right to amend course outlines occasionally and without notice. Courses offered by other delivery methods may vary from their individualized-study counterparts.

Opened in Revision 1