Abstract:
The quality of education is the main focus in providing higher education. However,
the quality of education cannot be measured solely by looking at the results of the final
exam, therefore, a timely graduation prediction system is needed that can assist lecturers
and faculty staff in providing supervision and support to students in achieving their goals.
Student graduation on time is one of the assessments or benchmarks in the higher
education accreditation process. Higher education accreditation assessment is carried
out by the National Accreditation Board for Higher Education. The higher the
accreditation value, the higher the level of quality of higher education. In this case,
technology can be used as a solution to develop an effective and efficient prediction
system. This study focuses on predicting student graduation by Data Mining using the
Support Vector Machine algorithm and the level of accuracy of the algorithm is measured
using the Confusion Matrix. This study used data from students of the Information
Systems Study Program, Mathematics, Statistics, Chemistry, Biology, and Physics of the
Faculty of Mathematics and Natural Sciences, University of Riau who had graduated
from 2014-2017 as many as 1045 data. Ability to classify student graduation predictions
with an accurate 99,04% data accuracy rate.
