In the entire data mining process, the data cleaning process
is utilized in order to eliminate irrelevant items. The
discovery of patterns will be only useful if the data
represented in files offer a real representation of the
enrolment process and the actions or decisions taken by the
past student [16]. Here we study the enrollment factors and
decisions taken by students for enrollment in a particular
institute. Subsequently if the student is enrolled in a
particular institute then he is considered as a loyal student.
Initially, the institute provided us a database equal to 19012.
After filtering process of the data 2106 records were left.
The data supplied is from a particular institute, we cannot
mention its name, but it is from the Arabic region. The main
objective of our research is to discover patterns that will be
used to predict a loyal or not loyal student previously to his
enrollment or not enrollment to a particular institute. By
taking information from other students with similar
information, in this sense, we can know the role of each
attribute and the implicit relations among them