Analisis Diskriminan Logistik untuk Klasifikasi Sekolah Standar Internasional Berdasarkan Status Afiliasi

Abstract

Affiliate program is one of the government's program to improve the quality of education in Indonesia. The main activity in this program is a partnership with schools in developed countries so that interactions can occur between the components of the school. Implementation of this program need support of the government and society. In order that all support can be utilized maximally, required methods to classify and predict an affiliation status of school. Logistic discriminant analysis is one of the alternatif method to classify which the variables is a combination of numerical and categorical. Logistic discriminant function that used to classify of schools into the particular affiliate status is obtained by logit model predicted as a result of final reduction models.. Through the partial parameter testing, variables that affect the status of affiliation is school type, education of headmaster, and amount of the monthly tuition of international class. The classification accuracy of the discriminant function that is equal to 66.46%.