Aplikasi Classification and Regression Tree (CART) dan Regresi Logistik Ordinal dalam Bidang Pendididikan

(Studi Kasus: Predikat Kelulusan Mahasiswa S1 Fakultas Matematika dan Ilmu Pengetahuan Alam Universitas Mulawarman)

Abstract

CART method is a nonparametric statistical methods which is for obtaining accurate data group in the classification analysis. CART main goal is to get an accurate data as a group identifier of a classification. CART can be applied in three main steps, namely the establishment of a classification tree, trimming the classification tree, and determination of optimal classification tree. Ordinal logistic regression is a statistical method for analysis response variables that have an ordinal scale consisting of three or more categories. Predictor variables that can be included in the model can be either continuous or categorical data consisting of two or more variables. This study wanted to know the classification results FMIPA UNMUL predicate graduation, the main factor that affect the predicate graduation FMIPA UNMUL who graduated in 2014, and a comparison of the accuracy of the classification results between CART and ordinal logistic regression. The results showed that gender (X₁), region origin (X₂), major (X₃), the status of secondary school (X₄), and duration of the study period (X₅) is the primary identifier graduation predicate FMIPA UNMUL, whereas gender (X₁ ) and duration of the study period (X₅) is a factor that affects the predicate graduation. Ordinal logistic regression model was able to predict with 65% accuracy, while the CART method has a predictive accuracy of 54.9%