KLASIFIKASI TINGKAT KESEJAHTERAAN KELUARGA DI KECAMATAN SIDEMEN MENGGUNAKAN BOOTSTRAP AGGREGATING (BAGGING) REGRESI LOGISTIK ORDINAL

Abstract

This research was conducted to determine the variables that have a significant impact on the stages of a well-off family in Sidemen Sub-district based on indicators obtained from the BKKBN and to classify the stages of a well-off family. This study used secondary data obtained from the stage of well-being data, Sidemen Sub-district, Karangasem Regency from BKKBN, totaling 1796 families. The method used is ordinal logistic regression and bagging ordinal logistic regression. Based on the logit regression model of ordinal logistic regression and ordinal logistic regression bagging, there are fourteen variables that have a significant effect on the dependent variable, namely marital status, type of insurance, age of head of household, occupation of head of household, having a source of income, eating a variety of food, having savings, accessing information from online media, families have ever recreated together, families have ever participated in social/community activities, the largest type of floor, main source of drinking water, ownership of a house/building, and children are still in school. The classification accuracy level in testing data using the ordinal logistic regression method was 79.4%, while the classification accuracy level using the bagging ordinal logistic regression method with 50,000 replications was 82.78%, so bagging showed an increase in classification by 3.38%.