Applicant Mapping of Predicting Seriousness in Choosing Private University using Self Organizing Map (SOM) Algorithm

The admission of new students is an initial effort made by PTS, but many applicants have less serious interest, so it has a negative impact on universities. Universities have to work very hard, to find prospective students, when the competition is getting tougher from time to time. Therefore we need a system that can predict, between truly serious and less serious applicants. Self Organizing Map (SOM) application, is a method of artificial neural network (ANN) that can be used to make predictions. Prediction of applicants is done by mapping the characteristics of the applicants, based on nine indicators obtained from the applicant's input data. The nine indicators consist of pathways of entry, distance from home to university, applicant's job, father's job, mother's occupation, parent's income, year of graduation, school of origin, and official services. The system with SOM prediction can do two mappings, namely serious registrants and less serious registrants. Comparison between predicted data and reality in the field, on the MAPE value or the average percentage error value each year. In 2017 it was 0.0065%, 2018 was 0.00067%, and in 2019 it was 0.0047%. From the results of predictive research using SOM, it can be seen that the trend pattern of registrant interest is the same as the data in the field which shows a decreasing graph from year to year.