Search for collections on Repository Universitas Sulawesi Barat

PENERAPAN METODE DBSCAN (Density-Based Spatial Clustering of Applications with Noise) UNTUK MENGELOMPOKKAN POTENSI LAHAN KELAPA DI PROVINSI SULAWESI BARAT. APPLICATION OF DBSCAN (DENSITY-BASED SPATIAL CLUSTERING OF APPLICATIONS WITH NOISE) METHOD TO GROUP COCONUT LAND POTENTIAL IN WEST SULAWESI PROVINCE

ARYA NUGRAHA, KALILI (2025) PENERAPAN METODE DBSCAN (Density-Based Spatial Clustering of Applications with Noise) UNTUK MENGELOMPOKKAN POTENSI LAHAN KELAPA DI PROVINSI SULAWESI BARAT. APPLICATION OF DBSCAN (DENSITY-BASED SPATIAL CLUSTERING OF APPLICATIONS WITH NOISE) METHOD TO GROUP COCONUT LAND POTENTIAL IN WEST SULAWESI PROVINCE. Diploma thesis, Universitas Sulawesi Barat.

[thumbnail of SKRIPSI ARYA NUGRAHA KALILI.pdf] Text
SKRIPSI ARYA NUGRAHA KALILI.pdf

Download (1MB)
[thumbnail of SKRIPSI ARYA NUGRAHA KALILI (pdf.io).pdf] Text
SKRIPSI ARYA NUGRAHA KALILI (pdf.io).pdf
Restricted to Repository staff only

Download (771kB)

Abstract

Provinsi Sulawesi Barat memiliki potensi besar dalam pengembangan komoditas kelapa, namun pemanfaatannya belum optimal karena kurangnya analisis spasial yang terstruktur. Penelitian ini bertujuan untuk mengelompokkan potensi lahan kelapa menggunakan algoritma Density-Based
Spatial Clustering of Applications with Noise (DBSCAN) sebagai pendekatan data mining untuk analisis spasial pertanian. Data yang digunakan merupakan data simulasi (dummy) yang menyerupai data dari Badan Pusat Statistik (BPS) dan Dinas Perkebunan Provinsi Sulawesi Barat, dengan variabel utama yaitu luas lahan, produksi, dan produktivitas. Setelah melalui tahapan data preprocessing dan normalisasi Min-Max, diperoleh parameter optimal ε = 0,767 dan MinPts = 5, dengan nilai Silhouette Coefficient sebesar 0,3101, yang menunjukkan kualitas klasterisasi cukup baik untuk data spasial dengan variasi kepadatan tinggi. Hasil analisis menghasilkan enam cluster utama dan satu kelompok noise, yang masing-masing mewakili tingkat potensi lahan dari tinggi hingga rendah. Visualisasi hasil pengelompokan melalui peta Geographic Information System (GIS) menunjukkan pola distribusi spasial yang konsisten antara produktivitas kelapa dan kondisi wilayah, di mana daerah dataran rendah memiliki potensi yang lebih tinggi dibandingkan daerah pegunungan. Temuan ini menegaskan bahwa algoritma DBSCAN efektif untuk mengidentifikasi pola alami data spasial tanpa menentukan jumlah cluster di awal, serta dapat mendukung pengambilan keputusan berbasis data bagi pemerintah daerah dalam pengembangan sektor perkebunan kelapa secara berkelanjutan di Provinsi Sulawesi Barat.
West Sulawesi Province has significant potential for coconut cultivation development; however, its utilization has not been optimal due to the lack of structured spatial analysis. This study aims to classify coconut land potential using the Density-Based Spatial Clustering of Applications with Noise (DBSCAN) algorithm as a data mining approach for agricultural spatial analysis. The data used are simulated (dummy) data modeled after datasets from the Central Bureau of Statistics (BPS) and the Plantation Agency of West Sulawesi Province, with key variables including land area, production, and productivity. After undergoing data preprocessing and Min-Max normalization, the optimal parameters were determined as ε = 0.767 and MinPts = 5, with a Silhouette Coefficient value of 0.3101, indicating moderate clustering quality for spatial data with high density variation. The analysis produced six main clusters and one noise group, each representing different levels of land potential from high to low. Visualization through a Geographic Information System (GIS) map revealed consistent spatial patterns between coconut productivity and topographical conditions, where lowland areas showed higher potential compared to mountainous regions. These findings confirm that the DBSCAN algorithm is effective in identifying the natural structure of spatial data without requiring a predefined number of clusters and can support datadriven decision-making for local governments in planning sustainable coconutplantation development across West Sulawesi Province.

Item Type: Thesis (Diploma)
Uncontrolled Keywords: DBSCAN, Clustering, Potensi Lahan Kelapa, Data Spasial, Sulawesi Barat, Silhouette Coefficient. DBSCAN, Clustering, Coconut Land Potential, Spatial Data, Silhouette Coefficient, West Sulawesi, Dummy Data.
Subjects: FAKULTAS TEKNIK > Informatika
Divisions: Fakultas Teknik
Depositing User: Chaeril Anwar
Date Deposited: 15 Jun 2026 04:48
Last Modified: 15 Jun 2026 04:48
URI: https://repository.unsulbar.ac.id/id/eprint/2899

Actions (login required)

View Item
View Item