Identifikasi Faktor-Faktor yang Mempengaruhi Status Anemia pada Ibu Hamil Menggunakan Metode CHAID dan Random Forest

Abstract

Maternal anemia is one of problem in society. The high rates of maternal anemia in line with lack of awareness to iron consumption. The solution for this problem is classifiying the factors associated with maternal anemia status. CHAID method (Chisquare Automatic Interaction Detection) is one of classification method which type categorical response variable. In this study, CHAID method applied to eight explanatory variable with repetition three times so that the resulting classification tree is different. Differences in classification result with the CHAID method lead to instability identifier variable therefore proceed with the random forest method to see identifier variable that stable. Random forest produced three explanatory variables sequentially stable are upper arm circumference (lila), frequency of prenatal care (anc), and maternal age clinical. Identifier variable be produced random forest method applied to the CHAID method to see its classification. CHAID method with random forest criteria produce that maternal anemia criteria if maternal upper arm circumference (lila) less than 23.5 cm and frequency of prenatal care (anc) less than four times, where as maternal normal if maternal upper arm circumference (lila) more than 23.5 cm and frequency of prenatal care (anc) more than one.