Analysis of Strategy for Targeted New Student Using K-Means Algorithm

Abstract

The admission of new students at Universitas Klabat is held every semester. Especially in Faculty of Computer Science which has two departments, namely Information Systems and Informatics that are growing. So that new student data continues to grow every year. The data obtained can be managed to produce important information as an institutional reference in making decisions, such as determining a promotion strategy. Through the right promotion strategy, Faculty can reduce costs for promotion. Therefore, this research was conducted to determine an effective and efficient promotion strategy to targeted new student and information can be spread on the right target. This study uses data from newly enrolled students consisting of 243 data using several attributes such as program study, gender, origin, religion, school, major, and promotion. We used K-Means algorithm, which is one of the non-hierarchical clustering data methods in classifying student data into several clusters based on data similarity, so that students who have the same characteristics are grouped into one cluster. We also used a statistical gap method to determine an optimum cluster. The cluster of students was classified into three clusters in the followings: cluster 0 produces 59 data, cluster 1 produces 94 data, and cluster 2 produces 90 data.