Automatic Quiz Generation Mechanism for Multiple Choices Question by Applying Ontological Data

Abstract

       Quiz generation system is a very important system in education. This helps to assess students' understanding of the lesson, and also allows users to easily and quickly create the test. However, the existing system of creating quizzes has limitation to the ability for creating a question, which is also entered by the user themself. As a result, it takes time to create a test for a long time. Therefore, this research aims to provide a mechanism for the creation of automatic quiz generation mechanism for multiple choices question by applying ontological information to assess the difficulty level of the questions. The hybrid similarity was measured using a combination of semantic similarity, semantic relatedness, and property values to determine the difficulty level of the question. Then, the proposed mechanism was implemented using the RDFaCE tool and PHP program to create an interface for user input and displaying results in the creation of quiz based on user-defined data. In addition, the proposed mechanism was validated to confirm the accuracy of the mechanism’s performance. By comparison, the difficulty score derived from the proposed mechanism and the item response theory. The evaluation results were consistent with 80 %. Therefore, the proposed automatic quiz generation mechanism can be applied as a tool to quickly and easily create quizzes and reduce the time of the work, including the creation of an increasing number of multiple choices question which are diverse. This method can also determine the difficulty level of the questions as required.