Klasifikasi Algoritma KNN dalam menentukan Penerima Bantuan Langsung Tunai

Abstract

Poverty is a condition where an individual or a group experiences economic incapacity to meet their basic needs, and this study involves broader aspects than just expenditures. The focus of this research is on Ciherang and Ciaro Villages, located in the Nagreg Subdistrict, Bandung Regency, which are areas with significant levels of poverty. This research responds to poverty issues by utilizing the K-Nearest Neighbor (KNN) Algorithm in the data mining classification process. KNN considers the proximity of a new object to its nearest neighbors, and as a supervised learning algorithm, KNN requires target information or classes in the analyzed dataset. The aim of this research is to provide information regarding the classification of recipients of Direct Cash Assistance (BLT) in Ciherang and Ciaro Villages. The research results present data on criteria for determining whether recipients are considered eligible or ineligible for BLT, with an accuracy rate reaching 81.56%. Additionally, the performance of this algorithm is demonstrated through true recall values for both eligible and ineligible recipients, with recall for true ineligible recipients at 88.43%, recall for true eligible recipients at 74.80%, precision for eligible recipients at 86.79%, and precision for ineligible recipients at 77.54%. These findings provide a basis for more accurate decision-making in determining BLT recipients in both villages. This can contribute to the design of more targeted and effective social policies in reducing the impact of poverty, providing in-depth insights into the characteristics of BLT recipients, and demonstrating the relevance and efficiency of the KNN algorithm in addressing complex social issues.