Applying Predictive Analytics using Clickstream Data for Improving the Students Performance

Abstract

Student performance analysis is an essential aspect of educational institutions. In recent years, machine learning (ML) has emerged as a promising tool to analyze student performance. To predict learning abilities of students and prescribing them a personalized learning curriculum, it is necessary to estimate their behavior to know about their weaknesses, strengths and help institutions to improve enrollment and retention. If it is possible for the teachers to predict in advance and prescribe ways to the at-risk and dropout students, they can plan more effectively to help them. We are describing in this paper various intelligent tutoring systems with Educational Data Mining, Predictive Learning Analytics, prediction of at-risk students at an earlier basis and how this prediction task is done. Predictive Analytics can also offer insights to help students to make informed decisions about each individual student to improve outcomes by understanding what drives each student behavior and how much the institution can create intensional, specific plans that will positively impact students.