Automatic generation of quiz questions using Freebase

Abstract

In this thesis an algorithm is designed to generate questions and answers au- tomatically and randomly using Freebase to be used in an online knowledge game, called Game of Gums. In this game, a player chooses a topic based on his/her interest and answers to some questions in order to go further steps in the game. Freebase is a knowledge base contains millions of entities in different domains. By accessing Freebase data by using Freebase APIs, we can read Freebase data and use entities and their properties for generating questions and answers. Questions in this game are in different category levels of : \Easy", \Medium", and \Hard". These categories are created based on the score of entity that are assigned to that during indexing. The score shows the popularity and number of incoming and/or outgoing links to each entity. Furthermore, a Java program was implemented for each question and corresponding answers using NetBeans IDE. Finally, surveys are designed and have been sent to interested volunteers in each topic to evaluate quiz questions. The results of these studies were discussed afterward. The results show that most of the respondents could answer easy questions while least of them answer hard ones. All these steps are explained in this thesis in great detail.