Title :
Towards automated assessment of engineering assignments
Author :
Quah, Jon Tong-Seng ; Lim, Luo-Ren ; Budi, Hendri ; Lua, Kim-Teng
Abstract :
This paper reports an automated engineering assignments marking system using support vector machines (SVMs). A typical engineering assignment consists of more than just text. It may also contain mathematical equations, pictures, diagrams, charts, algorithms, or even programming source codes. These elements have to be taken into consideration in the marking process. The automated marking process is more consistent. The system learns how to mark based on grades given for the first few scripts. The system would learn the marking scheme and mark the subsequent scripts automatically. A prototype system to mark equations and short answers is prototyped and evaluated. This system forms the foundation to be expanded into a full fledge automated assessment system.
Keywords :
educational technology; engineering education; support vector machines; automated engineering assignments marking system; educational technology; engineering education; learning systems; mathematical equations; programming source codes; support vector machines; Circuit testing; Design engineering; Educational technology; Equations; Mathematical programming; Neural networks; Programming profession; Prototypes; Support vector machines; Systems engineering and theory; educational technology; engineering education; learning systems; support vector machine;
Conference_Titel :
Neural Networks, 2009. IJCNN 2009. International Joint Conference on
Conference_Location :
Atlanta, GA
Print_ISBN :
978-1-4244-3548-7
Electronic_ISBN :
1098-7576
DOI :
10.1109/IJCNN.2009.5178782