A COMPARISON OF FUZZY TIME SERIES AND LEAST-SQUARE METHOD IN FORECASTING STUDENTS’ ENROLMENT

Abstract

Enrolment forecasting, which provides information for decision making and budget planning, is important in many ways to higher education. Because of its importance, researchers have proposed many forecasting methods to improve accuracy. Different methods such as genetic algorithm, least square that are used to forecast enrolment of student do not give relatively accurate results. However, obtaining accuracy is not an easy task, as many factors have impacts on enrolment numbers. In this work, a fuzzy time series was developed for efficient enrolment forecasting. The model is made up of four steps which are definition of the universe of discourse and intervals, fuzzification of historical data, establishment of fuzzy relationships and enrolment forecast. The max-min operator was used as universe of discourse and we compared our proposed method with the existing linear regression method. The historical enrolment figures of the University of Agriculture, Abeokuta were used as a data set for testing and were implemented using Visual Basic. The forecasting result of the fuzzy time series method is compared with that of the existing least square method, the fuzzy time series method produces the smallest values of the mean square error (MSE) as compared with the least square method. The application was also used to predict students’ enrolment for the next five years. The proposed method was found to obtain more accurate forecasting results than the existing method.