TY - JOUR
T1 - A novel approach to generate distractors for Multiple Choice Questions
AU - Kumar, Archana Praveen
AU - Nayak, Ashalatha
AU - Manjula Shenoy, K.
AU - Goyal, Shashank
AU - Chaitanya, null
N1 - Publisher Copyright:
© 2023 Elsevier Ltd
PY - 2023/9/1
Y1 - 2023/9/1
N2 - Multiple Choice Questions (MCQs) have been predominantly used as an assessment tool in the educational domain. The MCQ comprises a question called ‘Stem’, one correct answer called ‘Key’, and the incorrect options called ‘Distractors’. Identifying distractors is an essential step in MCQ construction because distractors need to be misleading and plausibly incorrect. Therefore the manual construction of MCQ is error-prone, and requires cumbersome efforts. Hence existing works have focused on automatic generation of MCQs but primarily towards vocabulary assessment. However, very few works inclined towards the technical domain have failed to analyze the plausibility of distractors. In this context, the proposed research DIstractor GENeration (DIGEN) is targeted to generate distractors for the MCQ in the technical domain. Hence, the novel contribution here is DIGEN takes unstructured text as well as multiple-choice questions with key as mandatory source along with ontology which may be an optional source to generate distractors automatically in the technical domain. Distractors generated have been evaluated based on Item Response Theory, which shows promising results.
AB - Multiple Choice Questions (MCQs) have been predominantly used as an assessment tool in the educational domain. The MCQ comprises a question called ‘Stem’, one correct answer called ‘Key’, and the incorrect options called ‘Distractors’. Identifying distractors is an essential step in MCQ construction because distractors need to be misleading and plausibly incorrect. Therefore the manual construction of MCQ is error-prone, and requires cumbersome efforts. Hence existing works have focused on automatic generation of MCQs but primarily towards vocabulary assessment. However, very few works inclined towards the technical domain have failed to analyze the plausibility of distractors. In this context, the proposed research DIstractor GENeration (DIGEN) is targeted to generate distractors for the MCQ in the technical domain. Hence, the novel contribution here is DIGEN takes unstructured text as well as multiple-choice questions with key as mandatory source along with ontology which may be an optional source to generate distractors automatically in the technical domain. Distractors generated have been evaluated based on Item Response Theory, which shows promising results.
UR - http://www.scopus.com/inward/record.url?scp=85153240112&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85153240112&partnerID=8YFLogxK
U2 - 10.1016/j.eswa.2023.120022
DO - 10.1016/j.eswa.2023.120022
M3 - Article
AN - SCOPUS:85153240112
SN - 0957-4174
VL - 225
JO - Expert Systems with Applications
JF - Expert Systems with Applications
M1 - 120022
ER -