Automatic Concepts Classification based on Bloom's Taxonomy using Text Analysis and the Naive Bayes Classifier Method

摘要

This paper aims to add Bloom's Taxonomy levels as tags to the contents (e.g. concepts) of any given text-book which is written in formal English and given as a course material. Bloom's Taxonomy levels defines concepts and knowledge of learning as six levels. Preparing the material of any course based on these six could help the students to better understand the course's concepts and their interrelationships. However, the relations between concepts are highly sophisticated and require a human judgment. A set of methods have been proposed to extract the relations among concepts. We use the naive Bayes classifier which is the best known and most successful classification technique in Machine Learning (Mahesh Kini M et al., 2015). This work presents a naive classifier method which identifies the Bloom's Taxonomy levels in text paragraphs based on some rules in the training set. We evaluate and validate the proposed method on a text-book. By utilizing the concepts of computer science for determining its knowledge domain. As a result of the proposed method achieves an accuracy of average 70-85%, which is significantly high. Furthermore, we show that taking Bloom's Taxonomy levels into account in course design is valuable and our method can be used to achieve.