Automatic short answer grading using rough concept clusters Online publication date: Wed, 06-Nov-2019
by Udit Kr. Chakraborty; Debanjan Konar; Samir Roy; Sankhayan Choudhury
International Journal of Advanced Intelligence Paradigms (IJAIP), Vol. 14, No. 3/4, 2019
Abstract: Evaluation of text-based answers has stayed as a challenge for researchers in recent years and with the growing acceptance of e-learning system, a solution needs to be achieved fast. While assessing the knowledge content, correctness of expression and linguistic patterns are complex issues in themselves, a smaller answer may be evaluated using keyword matching only. The work proposed in this paper is aimed at evaluating smaller text answers, no longer than a single sentence using keyword matching. The proposed method agglomerates keywords from a group of model answers forming clusters of words. The evaluation process thereafter exploits the inherent roughness of the keyword clusters to evaluate a learners' response through comparison and keyword matching. The novelty in the proposed system lies in the usage of fuzzy membership functions along with rough set theory to evaluate the answers. Rigorous tests have been conducted on dataset built for the purpose returned good correlation values with the average of two human evaluators. The proposed system also fares better than latent semantic analysis (LSA) based and link grammar-based evaluation systems.
Existing subscribers:
Go to Inderscience Online Journals to access the Full Text of this article.
If you are not a subscriber and you just want to read the full contents of this article, buy online access here.Complimentary Subscribers, Editors or Members of the Editorial Board of the International Journal of Advanced Intelligence Paradigms (IJAIP):
Login with your Inderscience username and password:
Want to subscribe?
A subscription gives you complete access to all articles in the current issue, as well as to all articles in the previous three years (where applicable). See our Orders page to subscribe.
If you still need assistance, please email subs@inderscience.com