PENERAPAN ANALISIS DISKRIMINAN LINIER ROBUST DALAM PENGKLASIFIKASIAN INDEKS KEPEDULIAN TERHADAP ISU KEPENDUDUKAN (IKIK)

Abstract

Abstract

Robust linear discriminant analysis is used to classify data that contains outlier by replacing classical parameters in linear discriminant analysis with robust parameters. This study aims to classify the Index of Concern for Population Issues (IKIK) of 34 provinces in 2020 into two categories namely target fulfilled IKIK and target not fulfilled IKIK using robust linear discriminant analysis. The independen variabels used are quantity dimensions (X_1), quality dimensions (X_2), mobility dimensions (X_3), and environment dimensions (X_4). The results obtained are 17 provinces were categorized as target fulfilled IKIK, 17 provinces as target not fulfilled IKIK. There are 2 robust discriminant functions formed, each for target fulfilled and target not fulfilled IKIK. The accuracy of the robust linear discriminant functions formed is 97,06%, the APER value of the discriminant functions is 2,94% and the PressQ value = 30,11 is greater than the value of (Chi_(3,0.05)^2) = 7.81. This shows that the discriminant functions can classify observations accurately.

Keywords: classification, Index of Concern for Population Issues, robust discriminant analysis