ABSTRACT
The main goal of all educational institutions is for students to do well in school. To help students do better in school, it is important to know how determined they are and to be able to predict their academic success. Knowing about a student’s academic performance ahead of time is helpful for a number of reasons. It can help universities come up with plans to keep or improve students’ grades; it can help management make better decisions about who to admit; it can help universities grow and improve in a way that won’t hurt them in the long run; and it can help students keep or improve their grades. This article presents a machine learning based framework for academic performance prediction. Data acquisition is the first step for student performance classification and prediction. Data is cleaned to remove noise or missing data. Useful features are selected using CFS technique. Classification model consist of SVM and NB techniques.
