A Conceptual Framework to Aid Attribute Selection in Machine Learning Student Performance Prediction Models

One of the important key applications learning analytics is offering an opportunity to institutions track student’s academic activities and provide them with real-time adaptive consultations if student performance diverts towards inadequate outcome. Still, numerous barriers exist while developing implementing such kind applications. Machine algorithms emerge as useful tools endorse by building models capable forecasting final outcome students based on their available attributes. The machine algorithm’s demotes using entire attributes thus a vigilant selection predicting boosts produced model. Though, several constructive techniques facilitate identify subset productive attributes, however, challenging task evaluate prediction are meaningful, explicit, controllable students. This paper reviews existing literature come up used in models. We propose conceptual framework which demonstrates classification either latent or dynamic. may appear significant but not able control these attribute, other hand, has command restrain dynamic Each major class further categorized present researchers pick for model development.