Mining students outcomes: An empirical study

Abstract

The purpose of the research presented in this paper is to extract students' individualized learning achievements from course information and assessments results for student groups and predict expected performance in future courses based on existing achievements in a set of student outcomes at the program level. Data mining techniques are used to process course information and extract students' achievements in a set of student outcomes. Two prediction algorithms namely single linear regression and multiple linear regression are applied to determine students' expected performance in future courses. Specialized programs have been designed, implemented and processed in stages to achieve the results as described in this paper. The results show that by using data mining techniques individualized students' performance information can successfully be extracted to provide formative feedback. This information is further processed by applying prediction algorithms to determine expected students' performance in future courses depending on the available historical records for courses.