Klasifikasi dan Klastering Mahasiswa Informatika Universitas Muhammadiyah Surakarta

Abstract

Data in an organization which are currently increasingly more and accumulate, will not lead to the use of data become optimal. Informatics Department in UMS that has been established since 2007 is one of the study programs that have a large data. The amount of this data will only be a pile of data if it is not processed into strategic information with certain methods, such as classification and clustering. This study was done to take advantage of the abundant data as a source of strategic information for the department to classify the students’ length of study and the students’ degree of excellence and also to cluster them using data mining techniques. Students’ classification was done by using Decision Tree from 223 graduated students’ data, while the clustering was conducted by using K-Means algorithm from 209 active students’ data. Attributes used in this study consists of high school majors, gender, high school region, the average number of credits hour per semester, and students’ participation as an assistant which were set as an independent variable. While the length of study and the degree of excellence were set as the dependent variable. Informatics students classification and clustering shows that the most significant variable influencing on the length of study is the average of credit hours taken per semester by students, while the variables that most influence on students’ degree of excellence is student participation as an assistant. The result interprets that the variables that need to be used as consideration for the department to obtain the effective rate of the length of study is the average of credit hours, while the variable as consideration to obtain the maximum degree of excellence is the student participation as an assistant.