Predicting students’ academic grade using machine learning algorithms with hybrid feature selection approach

Data plays an important role where any prediction is to be made. Due to the advancement in technology there have been significant ways of collecting students’ data. Educators use lots of supportive digital system to teach and collect data, such as Learning Management System, Student Information system, etc. The data collection through these system is huge and nevertheless it can be called as a big data. Lots of research is being done in how to improve the students’ performance and overall educational performance and to tap key areas that may help in student progress. Traditional analysis of data includes sampling, whether it be a climate prediction or any performance prediction. In this research, the data collected through classroom observation, academic performance or learning management systems is used to predict the performance of the students. Lots of factors were present in the collected data but the main factors that add relevance for appropriate prediction are only few. This paper proposes how more than one feature selection algorithm results can be combined to improve the predictions with machine learning algorithm.