IMPLEMENTASI ITERATIVE DICHOTOMISER 3 PADA DATA KELULUSAN MAHASISWA S1 DI UNIVERSITAS SEBELAS MARET

Thanh Thi Bi Dan, Sari Widya Sihwi, Rini Anggrainingsih

Abstract

University of Sebelas Maret (UNS) selects new students every year. The selection process usually appears in three ways, known as National Selection for University Entry (SNMPTN), Local Selection for University Entry (SBMPTN), and New Students Admission Selection (SPMB). Every year UNS also passes students through four periods of graduation. From this process, UNS has graduation data that increases every period. These data has not yet been used efectively and left to accumulate unorderly. Necessary technique is needed to process the data that can give standardized and satisfying results, thus by using data mining indeed can solve the problem.

In this research, the algorithm used is Iterative Dichotomiser 3 (ID3). ID3 algorithm is a method of learning that will build a decision tree that becomes model in finding solution of the problem. This algorithm in a greedily procedure searches  every possibility of the decision tree. ID3 can be used to process graduation data of UNS bachelor’s degree students that can give an significant information categorized in a 12 class structure bases on grade point average and graduation time. This group designed based on attribute gender, faculty, high school origin, enrollment process, parents’ salary, and parents’ job and this information can be used to determine new students acceptance in University.

From the tree model formed, then it tested by using measuring precision and recall. This testing were done by using data testing. Form the overall data 1066, used 95% for data training. Data testing used 5% from the overall data that is tested ten times randomly. From the testing result, it can be concluded that the implementation of ID3 algorithm on graduation data of UNS bachelor’s degree students has average maximum precision 63.96% and average maximum recall 62.47%  of data testing in a ten times testing.

Keywords

Classification, Data mining, Decision tree, Iterative Dichotomiser 3.

Refbacks

  • There are currently no refbacks.