A Study of Assessment Evaluation Schemes using Pattern Mining

Abstract

Learning is the main aspect of getting educated. Therefore, the enhancements in learning techniques is the question, that how and what aspects should be followed to make the learning techniques better and usable for various educational institutions. Hence, a study is carried out which assesses the student on the basis of what the student has learnt during the study cycle. This is one of the most important aspects of education system as it determines whether the technique of conducting exams is a correct approach to measure a learner’s abilities or not. It is an on-going debate between different school of thoughts whether conducting exams is the best way to measure the learners’ abilities or not. The formula for creating a perfectly balanced exam depends on a number of factors. Are the questions relevant to what was being taught to the students? Is the language used for questions clear and easy to understand? Another factor that can be a competitor in creating a good exam is the sequence of questions based on the level of answers. We study that the order of difficulty will help the students to determine overall approach to the questions. If a student faces difficult questions in the beginning and is unable to answer them, then student will be disappointed and may leave rest of the paper. Another scenario is that a student takes 2/3rd of time in completing first half of exam. Looking at students pace,students can panic and try to attempt other questions in hurry which will result in highly increased chances of mistakes in other questions as well. For this thesis, we perform analysis of the effects of question sequences based on varying difficulty patterns. To make it more reliable, we will be conducting the study using an android assessment application and apply Apriori’s algorithm and Frequent Pattern Mining to generate results.