APPLICATION OF THE APRIORI ALGORITHM TO DETERMINE THE COMBINATION OF POVERTY INDICATORS

Abstract

Poverty is a society that has not been solved until now. The decline in poverty in Laweyan District from 2000 to 2013 was 5.71%, among the five lowest in the reduction in the percentage of poverty in Central Java Province. The problem of poverty is very complex, and the differences in regional characteristics, as well as the techniques used, also influence the indicators of the causes of poverty and the formulation of policies for poverty alleviation. This study uses Principal Component Analysis as part of data preprocessing, followed by applying association rules with the Apriori Algorithm to explore the relationship pattern of poverty indicators. Based on the research that has been conducted on the poverty dataset, which consists of 46 attributes, it is found that the attributes that have passed the preprocessing data are six attributes, namely the Poor Population, ADHB in the Communication Sector, ADHB in the Mining and Excavation Sector, ADHB in the Agriculture and Food Crops Sector, ADHB in the Plantation Sector. and unemployment. These six attributes are transformed into Ascending, Fixed, and Descending categorical data. The fuzzification process for the increase and decrease categories uses the shoulder-type triangle membership function. Applying the Apriori Algorithm to the poverty dataset with a minimum support of 0.4 and a minimum confidence of 0.8 produces 38 rules that show the relationship between indicators and poverty and 134 rules that show the relationship pattern between indicators.