Student Prediction of Drop Out Using Extreme Learning Machine (ELM) Algorithm

Abstract

The purpose of this study was to predict students dropping out of the Education Management Doctoral Program of FKIP Mulawarman University and to evaluate the Extreme Learning Machine in predicting student dropouts. This research uses the Extreme Learning Machine algorithm, the feedforward neural network learning method and the Support Vector Machine algorithm for comparison of the level of accuracy using the same data. The data used is as much as 110 data according to the number of students from the class of 2012 to 2018, the data is taken from the SIA Education Management Study Program of the Mulawarman University Doctoral Program and then processed. In this case, how to predict student dropouts using the variable Gender, Semester 3 IP Value, Working Status, Family Status, Age, and using two DO and NON DO Classes? And calculating the accuracy value using a confusion matrix ?. From the results of this study, it can be concluded that students drop out in the Educational Management of the FKIP Mulawarman University Doctoral Program can be predicted by Extreme Learning Machine using the training value obtained from semester 3 of the 2012-2018 class. From the results of testing the predictive accuracy of the Extreme Learning Machine is 72 %.