Churn Prediction on Higher Education Data with Fuzzy Logic Algorithm

Abstract

Colleges are an optional final stage of formal education. However, with time, the Management Section finds the fact that the student churn rate in the university scope becomes a problem. The purpose of this research is to predict whether the student will be churn or loyal in the future, the data will be taken from 2014. The analytical technique used in this research process is the Fuzzy method of C-Means. At a university, the variable that will be tested is the Length of students who are currently attending higher education at the University or campus, when the last student pays enrollment, the last payment period, and the student's total payment. The data used as many as 100 datasets were collected in this research, starting from 2014 to 2020. Of the 100 Datasets of informatics engineering, students gained 92% of loyal students, and 8% of students predicted to churn.