ANALISA KEPENTINGAN ATRIBUT DATA PADA KLASIFIKASI HEREGISTRASI MAHASISWA STMIK WIDYA PRATAMA

Abstract

Students are the most valuable assets in a private tertiary institution. Because most of private tertiary institution revenue and operational costs are obtained from students. The number of students who carry out re-registration will obviously be a breath of fresh air for the institution. In the last 5 years it was noted that around 20% of prospective STMIK Widya Pratama student did not register. The latest data on 31 August 2018 recorded that 32.7% of registrants had not registered. The decline in the number of students can affect the financial stability of institutions primarily private colleges. Analysis of the best algorithm for the classification of student registration was carried out and proved that the decision tree C45 is the algorithm with the best accuracy.
Early knowledge of prospective students who might not be able to register can become an institution's reference to take action to retain students. Neatly arranged student data can be used by management to analyze the
characteristics and causes of students not to register. This research will
conduct an analysis of all existing data and attribute data. The method used in weighting is information gain which has been proven to be able to handle datasets with many types of attributes. The results of this study conclude that the work attributes of parents are the attributes with the highest level of importance. While the civil status attribute is the attribute with the lowest level of importance.

Keywords: student registration, information gain, acceptance of new students