Educational Data Mining for Predicting Students’ Academic Performance

Abstract

Educational Data Mining is an emerging trend that is used for automatically extracting purposeful information from collected data especially in the field of education for predicting future educational achievement or problems. Worldwide higher education institutes have huge number of students’ data; Data Mining has a potential to use this data for many purposes. Today’s impact of using ICT especially Social Media is taken as an advantage in educational process. This study focuses on identification of various factors which impact the performance of students’ learning during academic process, as well as how using Social Media can impact their future achievement. We also want to see how far we can predict about future performance on the bases of this information. This research is quantitative in nature, through purposive sampling students (N=200, size of cohort) of undergraduate’s program were selected as respondent from IQRA University of Karachi. The data was collected through structured Questionnaires. For data analysis and prediction Naive Bayes algorithms were implemented through Rapid Miner. On the basis of the responses received through questionnaire and their current Grade Point Average (GPA) prediction were made for their future performance. Results indicated that Social Media use for educational purpose has positive impact on students’ academic performance. Several other factors such as mother’s qualification, father’s qualification and family income also had an impact on students’ performance. This research will provide a guide line for stakeholders for predicting future trends and implication on achievement of students and developing proper planning and guidelines for them. This study can spot out those students who require special attention and will help in minimizing the failure by facilitating them to perform well.