NASRULLAH, Muhammad and Surarso, Bayu and Nurhayati, Oky Dwi (2024) ANALISIS ALGORITMA NAÏVE BAYES DAN K-NEAREST NEIGHBOR UNTUK MENDUKUNG PEMBERDAYAAN NELAYAN. Masters thesis, UNIVERSITAS DIPONEGORO.

	Text COVER DST-1.pdf Download (330kB)
	Text COVER DST.pdf Restricted to Repository staff only Download (1MB)
	Text BAB I.pdf Download (386kB)
	Text BAB II.pdf Download (733kB)
	Text BAB III.pdf Restricted to Repository staff only Download (911kB)
	Text BAB IV.pdf Restricted to Repository staff only Download (2MB)
	Text BAB V.pdf Restricted to Repository staff only Download (421kB)
	Text DAFTAR PUSTAKA.pdf Download (332kB)
	Text LAMPIRAN.pdf Restricted to Repository staff only Download (439kB)

Abstract

Pemberian bantuan kepada nelayan adalah salah satu upaya pemerintah untuk mendukung pemberdayaan nelayan. Klasifikasi kelayakan pemberian bantuan untuk nelayan ini penting dilakukan agar bantuan yang diberikan tepat sasaran dan resiko pemberian bantuan yang tidak tepat sasaran dapat dikurangi. Dalam hal ini, bagaimana memanfaatkan data mining dengan algoritma Naïve Bayes dan K-Nearest Neighbor untuk membangun model klasifikasi serta mengevaluasi dan membandingkan kinerja kedua model dalam melakukan klasifikasi nelayan yang layak sebagai penerima bantuan. Penelitian ini bertujuan untuk mengevaluasi dan membandingkan efektifitas algoritma Naïve Bayes dan K-Nearest Neighbor untuk mengklasifikasi kelayakan penerima bantuan nelayan yang relevan. Penelitian ini memanfaatkan metode data mining dengan menggunakan algoritma Naïve Bayes dan K-Nearest Negihbor serta menggunakan teknik imbalence dataset yaitu Synthetic Minority Over-sampling Technique (SMOTE) untuk melakukan klasifikasi kelayakan penerima bantuan nelayan. Dari hasil percobaan pada penelitian ini menunjukkan bahwa Naïve Bayes tanpa SMOTE mencapai akurasi 97.01%, presisi 94.16%, recall 96.67%, dan F1-score 95.39%, sedangkan KNN tanpa SMOTE mencapai akurasi 94.04%, presisi 94.53%, recall 86.00%, dan F1-score 90.06%. Implementasi SMOTE meningkatkan kinerja kedua algoritma, dengan Naïve Bayes mencapai akurasi 98.33%, presisi 96.86%, recall 100.00%, dan F1-score 98.49%, serta KNN mencapai akurasi 96.90%, presisi 97.72%, recall 96.19%, dan F1-score 96.94%. Naïve Bayes dengan SMOTE menunjukkan kinerja yang lebih baik dalam mengatasi ketidakseimbangan data dan memastikan identifikasi yang akurat terhadap nelayan yang layak menerima bantuan.
Kata Kunci: Naïve Bayes, K-Nearest Neighbor, Classification, Synthetic Minority Over-sampling Technique, SMOTE, Inbalance Dataset, Akurasi, Presisi, Recall, F1-score.

Providing assistance to fishermen is one of the government's efforts to support the empowerment of fishermen. Accurately classifying the eligibility for this assistance is crucial to ensure that the aid reaches the intended recipients, thus minimizing the risk of misallocation. This study focuses on leveraging data mining techniques through the Naïve Bayes and K-Nearest Neighbor algorithms to develop a classification model and to evaluate and compare the performance of these models in identifying fishermen eligible for assistance. The objective of this research is to assess and compare the effectiveness of the Naïve Bayes and K-Nearest Neighbor algorithms in classifying eligible fishermen for assistance. The study employs data mining methods that incorporate both the Naïve Bayes and K-Nearest Neighbor algorithms, utilizing the Synthetic Minority Over-sampling Technique (SMOTE) to address the issue of imbalanced datasets in the classification process. The findings from the experiments indicate that Naïve Bayes without SMOTE achieved an accuracy of 97.01%, precision of 94.16%, recall of 96.67%, and an F1-score of 95.39%. In contrast, KNN without SMOTE recorded an accuracy of 94.04%, precision of 94.53%, recall of 86.00%, and an F1-score of 90.06%. The application of SMOTE enhanced the performance of both algorithms, with Naïve Bayes attaining an accuracy of 98.33%, precision of 96.86%, recall of 100.00%, and an F1-score of 98.49%. Meanwhile, KNN reached an accuracy of 96.90%, precision of 97.72%, recall of 96.19%, and an F1-score of 96.94%. The Naïve Bayes algorithm combined with SMOTE demonstrated superior performance in addressing data imbalance and accurately identifying fishermen who qualify for assistance.
Keywords: Naïve Bayes, K-Nearest Neighbor, Classification, Synthetic Minority Over-sampling Technique, SMOTE, Inbalance Dataset, Accuracy, Precision, Recall, F1-score.

Actions (login required)

View Item