Perbandingan Algoritma K-Nearest Neighbors dan Naïve Bayes dalam Penentuan Penerima Bantuan di Desa Banyuputih Kidul

Thoriq Wahyu Hidayatullah
Ulya Anisatur Rosyidah
Nur Qodariyah Fitriyah

Jurnal Penelitian Teknologi Informasi dan Sains
Vol. 4 No. 1 (2026): : JURNAL PENELITIAN TEKNOLOGI INFORMASI DAN SAINS (JPTIS)

The distribution of social assistance represents a key government strategy to enhance the welfare of low-income communities. Nevertheless, its implementation frequently faces challenges related to inaccurate targeting, often caused by uneven data collection and subjective decision-making processes in identifying eligible beneficiaries. This study aims to compare the performance of the K-Nearest Neighbors (KNN) and Naïve Bayes algorithms in determining eligibility for social assistance recipients in Banyuputih Kidul Village. Both models were evaluated using a confusion matrix with performance indicators including accuracy, precision, recall, and F1-score. The findings reveal that the KNN algorithm outperformed Naïve Bayes in identifying recipients of the PKH social assistance program, achieving an evaluation score of 99%, compared to 86% for Naïve Bayes. These results indicate that KNN provides higher predictive reliability for eligibility classification. This research is expected to support the development of an objective, data-driven decision support system that can assist village governments in distributing social assistance more accurately and transparently.