The Distribution Pattern of New Students Admissions Using the K-Means Clustering Algorithm

Abstract

The characteristics of prospective new students who re-register can produce a distribution pattern of new student admissions according to the data obtained. The results of the distribution pattern can be used as material for decision making in determining the socialization and promotion of study programs at universities. To produce a pattern, it requires grouping based on data that has the same character. The k-means algorithm, which is an algorithm in data mining processing methods for clustering, is successful in classifying new student admissions by maximizing the similarity of characteristics between data in one cluster which is different from data in other clusters. This is evidenced by direct application based on data mining procedures to data on new student admissions for Engineering Faculty, Manado State University for 2018, 2019, and 2020 with 8 selected variables. After clustering using the RStudio application, Data 2018 generates 6 clusters, 2019 data generates 3 clusters, and 2020 data generates 4 clusters. There are some re-registration characteristics that are common to all three years, which can be read easily after clustering.