Asee peer logo

Bayesian Network Models for Student Knowledge Tracking in Large Classes

Download Paper |

Conference

2016 ASEE Annual Conference & Exposition

Location

New Orleans, Louisiana

Publication Date

June 26, 2016

Start Date

June 26, 2016

End Date

June 29, 2016

ISBN

978-0-692-68565-5

ISSN

2153-5965

Conference Session

NSF Grantees Poster Session I

Tagged Topic

NSF Grantees Poster Session

Page Count

10

DOI

10.18260/p.27282

Permanent URL

https://peer.asee.org/27282

Download Count

902

Paper Authors

biography

Chao Chen Department of Computer Science and Engineering, University of South Carolina

visit author page

Chao is a PhD student in the Department of Computer Science & Engineering at University of South Carolina. He is interested in applying machine learning algorithms and Bayesian statistics in social science study.

visit author page

biography

Ramin Madarshahian University of South Carolina

visit author page

Ramin is PhD student in Structural engineering in University of South Carolina. He also got Master of applied Science in Statistics at middle of his PhD program.
His main focus of research is uncertainty quantification in engineering and scientific problems.

visit author page

biography

Juan M Caicedo University of South Carolina

visit author page

Dr. Caicedo is an associate professor at the Department of Civil and Environmental Engineering at the University of South Carolina. His research interests are in structural dynamics, model updating and engineering education. He received his B.S. in Civil Engineering from the Universidad del Valle in Colombia, South America, and his M.Sc. and D.Sc. from Washington University in St. Louis. Dr. Caicedo's teaching interests include the development of critical thinking in undergraduate and graduate education. More information about Dr. Caicedo's research can be found online at http://sdii.ce.sc.edu

visit author page

biography

Charles E. Pierce University of South Carolina

visit author page

Dr. Pierce is a Bell South Teaching Fellow and Associate Professor in the Department of Civil and Environmental Engineering at the University of South Carolina. He is a member of the American Concrete Institute, American Society of Civil Engineers, and American Society for Engineering Education.

visit author page

biography

Gabriel Terejanu University of South Carolina

visit author page

Gabriel Terejanu is an Assistant Professor in the Department of Computer Science and Engineering at University of South Carolina. Previously he was a Postdoctoral Fellow at the Institute for Computational Engineering and Sciences at University of Texas at Austin. He holds Ph.D. in Computer Science and Engineering from University at Buffalo. He is currently working on the development of a comprehensive uncertainty quantification framework to accelerate the scientific discovering process and decision-making under uncertainty. Some projects currently supported by NSF and VP for Research include discovery of novel catalytic materials for biorefinery industry, modeling and prediction of naturally occurring carcinogenic toxins, and development of statistical models for tracking individual student knowledge.

visit author page

Download Paper |

Abstract

A computational tool is developed to track student knowledge of cornerstone engineering concepts in large classes. The model that is attempting to be achieved is that individual student knowledge is just a hypothesis/model that needs to be tested using data obtained from assessment instruments. This fits naturally with probabilistic methodologies such as Bayesian inference that formalizes the scientific method. This computational tool will go beyond grading, and will allow instructors to provide formative feedback with respect to the challenging concepts specific to each individual student. An initial pilot experiment has already been performed during the Fall 2015 Statics class. The data collected from 37 students over three sequential quizzes has been used to inform the development of Bayesian networks for knowledge tracking. The main challenge identified during this study relates to the instantiation of conditional probabilities of the questions’ answers given the knowledge of the concepts. While the conditional probability when knowing the concept can be easily constructed, the more interesting probability of answering incorrectly when not knowing the concepts is more challenging to define. This is due to the probability distribution for these cases being highly dependent on the question and the type of misconceptions that the students have at the time of assessment. To address this challenge we propose to learn these conditional probabilities directly from the data.

Chen, C., & Madarshahian, R., & Caicedo, J. M., & Pierce, C. E., & Terejanu, G. (2016, June), Bayesian Network Models for Student Knowledge Tracking in Large Classes Paper presented at 2016 ASEE Annual Conference & Exposition, New Orleans, Louisiana. 10.18260/p.27282

ASEE holds the copyright on this document. It may be read by the public free of charge. Authors may archive their work on personal websites or in institutional repositories with the following citation: © 2016 American Society for Engineering Education. Other scholars may excerpt or quote from these materials with the same citation. When excerpting or quoting from Conference Proceedings, authors should, in addition to noting the ASEE copyright, list all the original authors and their institutions and name the host city of the conference. - Last updated April 1, 2015