Abstract
STRATEGY TO IDENTIFY COGNITIVE LEVEL OF QUESTION USING BLOOM'S TAXONOMY
Asha B. Sattane*
ABSTRACT
An important component of question answering systems is question categorization. Questions are provided to fulfill learning objectives in the subject content learned by students. Challenging thing in question answering system is to prepare good quality questions. Quality questions are prepared by assigning cognitive level. Learning and assessment are the two sides of education system. Thus, Bloom's taxonomy is common reference point for it. Exam questions categorization presents a main challenge in categorizing short questions which will have small text. Short questions text are very sparse and far in terms of features also. In order to solve this issue, methodology is proposed to categorize exam questions automatically to the cognitive levels of Bloom's taxonomy. This provides a strategy based methodology using three machine learning classifiers. The classifiers adopted in this work are, Support Vector Machine (SVM),Naive Byes (NB), and k-Nearest Neighbour (k-NN) .The study found that applying feature selection methods, namely Chi-square, Mutual Information and Odd Ratio on question categorization not only make categorization more time efficient, but it also improves the categorization accuracy . Furthermore, it is discovered that combination of classifiers can be applied to categorize the question with feature selection methods.
[Full Text Article] [Download Certificate]