Penerapan Geospatial Artificial Intelligence Untuk Pengelompokan Provinsi Di Indonesia Berdasarkan Indikator Multidimensi Ketahanan Pangan

Fitri, Amelia Kurnia (2025) Penerapan Geospatial Artificial Intelligence Untuk Pengelompokan Provinsi Di Indonesia Berdasarkan Indikator Multidimensi Ketahanan Pangan. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 2043221009-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (8MB) | Request a copy

Abstract

Ketahanan pangan merupakan tantangan strategis Indonesia di tengah laju pertumbuhan penduduk pesat dan tekanan terhadap sistem pangan nasional. Meskipun negara agraris, realitas menunjukkan ketimpangan distribusi pangan yang signifikan di Indonesia, khususnya antara wilayah barat dan timur. Kondisi ini menunjukkan bahwa kebijakan yang sama untuk semua provinsi tidak lagi efektif. Karena itu, penelitian ini dilakukan untuk memetakan pola ketahanan pangan secara lebih presisi sebagai dasar penyusunan kebijakan yang lebih adaptif. Penelitian ini menerapkan pendekatan Geospatial Artificial Intelligence (GeoAI) dengan menggunakan data tahun 2024 yang memuat 14 variabel dari empat dimensi ketahanan pangan berdasarkan Food and Agriculture Organization (FAO). Dua metode reduksi dimensi digunakan sebelum proses klasterisasi, yakni Principal Component Analysis (PCA) dan t-distributed Stochastic Neighbor Embedding (t-SNE). Hasil reduksi tersebut dikelompokkan menggunakan algoritma Density-Based Spatial Clustering of Applications with Noise (DBSCAN) dan kemudian dibandingkan. Berdasarkan Silhouette Score, t-SNE–DBSCAN menghasilkan performa terbaik dengan nilai 0,559. Model terbaik ini mengidentifikasi tujuh klaster provinsi serta satu provinsi yang terdeteksi sebagai outlier karena memiliki karakteristik ekstrem. Seluruh hasil dianalisis dan divisualisasikan melalui dashboard interaktif berbasis Streamlit yang menampilkan peta sebaran dan profil masing-masing klaster. Disimpulkan bahwa pendekatan GeoAI efektif dalam mendeteksi anomali wilayah yang sering luput dari metode konvensional. Penelitian ini merekomendasikan pemanfaatan dashboard tersebut sebagai instrumen pendukung keputusan bagi pemerintah dalam merumuskan kebijakan adaptif, dengan memprioritaskan intervensi khusus bagi provinsi outlier yang rentan serta memperkuat stabilitas sistem bagi klaster yang telah mapan.
==================================================================================================================================
Food security is a strategic challenge for Indonesia amid rapid population growth and increasing pressure on the national food system. Although Indonesia is an agrarian country, significant disparities in food distribution remain, especially between the western and eastern regions. This condition indicates that a uniform policy for all provinces is no longer effective. Therefore, this study aims to map food security patterns more precisely as a basis for developing more adaptive policies. This research applies a Geospatial Artificial Intelligence (GeoAI) approach using 2024 data consisting of 14 variables across four Food and Agriculture Organization (FAO) food security dimensions. Two dimensionality reduction methods, Principal Component Analysis (PCA) and t-distributed Stochastic Neighbor Embedding (t-SNE), were employed before the clustering process. The reduced data were then grouped using the Density-Based Spatial Clustering of Applications with Noise (DBSCAN) algorithm and subsequently compared. Based on the Silhouette Score, the t-SNE–DBSCAN combination delivered the best performance with a score of 0,559. The optimal model identified seven provincial clusters and detected one provinces as outliers due to their extreme and uncommon characteristics. All results were analyzed and visualized through an interactive Streamlit dashboard displaying spatial distribution maps and detailed profiles of each cluster. The study concludes that the GeoAI approach is effective in detecting regional anomalies that are often overlooked by conventional methods. It recommends the utilization of the developed dashboard as a decision-support tool for policymakers, enabling more adaptive food security interventions, prioritizing vulnerable outlier provinces, and strengthening system stability in established clusters.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	DBSCAN, GeoAI, Ketahanan Pangan, PCA, t-SNE, DBSCAN, Food security, GeoAI, PCA, t-SNE
Subjects:	Q Science > QA Mathematics > QA336 Artificial Intelligence
Divisions:	Faculty of Vocational > 49501-Business Statistics
Depositing User:	Amelia Kurnia Fitri
Date Deposited:	19 Dec 2025 01:46
Last Modified:	19 Dec 2025 01:46
URI:	http://repository.its.ac.id/id/eprint/129050

Actions (login required)

View Item