ABSTRACT
Attrition or better known as student dismissal or drop out from
completing courses in higher learning
institutions is prevalent in higher learning institutions
in Malaysia and abroad. There are several reasons
attributed to the attrition in the context of student in hi
gher learning institutions. The degree of attrition
varies from one institution to another and it is cause for conce
rn as there will be a lot of wastage of
resources of academic and administrative besides the adverse e
ffect on the social aspect. In view of this,
minimizing the attrition rate is of paramount importance in ins
titutions. There have been numerous non-
technical approaches to address the issue, but they have not bee
n effective to predict at early stage the
likelihood of students dropping out from higher learning institut
ions. Technical approach such as data
mining has been used in predicting student attrition by some resea
rchers in their past research work.
However, not all prediction data mining techniques and othe
r relevant and significant factors attributed to
student attrition have been fully explored to address the issue.
As of result this, this study will focus on
using support vector machine model to predict probation status of st
udent in which in most cases will lead
to student’s dismissal. It will also examine relevant and othe
r factors that contribute to the attrition among
students in Malaysia. The result of the study is appealing as t
he support vector machine model achieves a
decent accuracy in prediction despite working on small size of
data set. With all this in place, higher
learning institutions in Malaysia can deploy the model in predict
ing probation status of student to minimize
student attrition