Development of a data mining software of higher educational data

Abstract

The purpose of this study is to develop a software which anlayzes the student data in Bogazici University, creates association rules and makes suggestions to the students about the course selection. The software works on two datasets. The first dataset accepts the grades as they are, so the grade set consists of 8 grades (AA, BA, BB, CB, CC, DC, DD and F). The second dataset groups similar grades and processes them together, so the gradeset consists of 4 grades such as (HIGH(AA-BA),MID(BB-CB,CC),LOW(DC-DD), FAIL(F)). The software consists of two main parts. The first part is the “Rule Creator Tool”, which analyzes the past data in the Bogazici University database and creates the rules. In rule creating part, the software uses association rules mining algorithm. The second part is the “User Tool”, which analyzes one student’s past data and finds the rules which are suitable for that student. After finding suitable rules, the software makes suggestions to the students about the course selection. The result of the study shows that the prediction accuracy may be increased about 2 times better than the naïve approach and 1.5 times better than the majority approach. This shows the data is suitable to run association rules mining algorithm. Keywords: Education, Association Rules Mining, Students, Advisors, Decision Support System (DSS), Grade Optimization, Course Selection.