Application of the Naïve Bayes Algorithm in Predicting Acceptance of Family Hope Program Assistance (PKH)

Abstract

PKH is one of the programs for providing conditional social assistance to Poor Families which are designated as recipient families, PKH benefits with the aim of relieving the economy of those who are less fortunate. However, in its implementation, especially in the community, this program has not run well and optimally due to the unevenness of the PKH program. This is of course very influential on the welfare of the community where the programs that should be received by the poor have not been well realized. The concept of data mining can be used to provide solutions to problems at hand, the Naïve Bayes Method is a method which can predict an opportunity from previous experience. The variables used include; Disability, children under five, school children and the elderly, where this variable is a requirement of a family as PKH recipients with additional criteria classified as a poor family with variables of occupation, income and residence status. Then the variable data is processed in the R language to determine the prediction results on the eligibility of PKH recipient families. The results of this study can be used as benchmarks and help in the optimization of the PKH program, from the experiments conducted, the results obtained were two households with id Test24 and Test44 which had the status of a recipient that was predicted to be non-recipient, with an Accuracy of 96%, 100% Sensitivity, and Specivicity 92%.