PEMBENTUKAN POHON KEPUTUSAN (DECISION TREE) MENGGUNAKAN METODE CHAID DAN CART PADA STUDI KASUS PERNIKAHAN USIA DINI

Abstract

Globally and Indonesia have a lot of child marriage. In East Kalimantan as much 440 child has marriage in 2017, and in juni 30th, 2018 has increase to 953 case. The risk involved to baby and mother like mortality, premature, stunting, miscarriage, mola hydatidosa, etc. This research aims to decision tree classification by CHAID and CART on Case Study of Early Marriage. Non-reactive research with cross sectional study, 1,221 population with 518 samples, the variables consist of early marriage, premarital childbearing, age at sexual debut, migration, residence, occupation, economic status, and education, using secondary data of SDKI 2017, data analysis univariate and bivariate (CHAID and CART). The result of CHAID classification accuracy training and testing data is 75.9% and 64.3% with one significant variable. CART classification result is 72.8% and 70.5% with 3 significant variables. CHAID accuracy on the work variable is 40.2% and 42.9%, education is 59.2% and 59.35. CART accuracy on employment variables is 40.5% and 46.9%, Economic Status is 29.4% and 30.8%, Education is 60.3% and 60.8%. CART was chosen as the better method to use, the girls with high impact to early marriage is age first intercourse at child and adolescent. Early marriage has an impact on work, economic status, and education. The suggestions from researchers are to increase prevention and program management efforts, expand information, and change family mindset related to early marriage.