Automatic Question Classification for Descriptive Answer Assessment Using Bloom’s Taxonomy

Abstract

Technology has set its traces in every walks of human knowledge development. The rapid development in technology impacts global education too. The growth of online education has not been without challenges which concern educating (training) and evaluating the knowledge of a learner. Prevailing systems in academic institutions evaluate the scores using online platforms for varied types of questions such as Multiple Choice Question (MCQ). Since these are pattern matching method, objective type question that needs one word as answers, analysis and validation need less effort. The same method is less feasible for any descriptive typescripts which have not yet been revolutionized to a greater extent. Therefore classifying the questions, identifying the right focus and extracting the meaning of the answers is one of the key challenges in assessment process. The prime focus of this research is to implement a system that undergoes question classification based on the knowledge dimension pointed in Bloom’s taxonomy. The system automatically evaluates the answers without any difference in evaluation that are likely to happen during manual evaluation which includes natural language processing and a DNN. The system aims to establish a method to automatically identify the levels of Bloom’s Taxonomy and evaluate the descriptive type of answers with the help of Natural Language Processing and appropriate intelligent system methodology that provides the legitimate result as well.