Nugroho, Salma Fitri and Luthfatul Amaliana, S.Si,. M.Si. (2024) Penerapan Algoritma Dbscan Dengan Reduksi Dimensi Pca Dan T-Sne Pada Pengelompokan Provinsi Di Indonesia Berdasarkan Indikator Ketahanan Pangan. Sarjana thesis, Universitas Brawijaya.
Abstract
Dalam melakukan analisis cluster, terdapat permasalahan umum yang sering terjadi seperti ‘curse of dimensionality’ pada ruang dimensi tinggi. Permasalahan ini dapat ditangani dengan melakukan reduksi dimensi sebelum dilakukan clustering. Kerawanan pangan merupakan salah satu isu global bagi berbagai negara, termasuk Indonesia. Berdasarkan informasi dari Badan Ketahanan Pangan, tingkat ketahanan pangan di Indonesia pada tahun 2022 berada di level 60,2. Posisi ini berada di bawah rata-rata global sebesar 62,2. Pada penelitian ini, dilakukan analisis cluster untuk memetakan provinsi di Indonesia berdasarkan 17 indikator ketahanan pangan menggunakan algoritma DBSCAN (Density-based Spatial Clustering of Application with Noise). Kemudian dilakukan perbandingan koefisien Silhouette pada data hasil reduksi PCA (Principal Component Analysis) dan data hasil reduksi t-SNE (t-distributed Stochastic Neighbor Embedding) untuk mendapatkan hasil cluster terbaik. Terdapat 9 cluster yang terbentuk tanpa adanya noise dari data hasil reduksi t-SNE dengan koefisien Silhouette sebesar 0,7761 yang termasuk kedalam kategori kuat.
English Abstract
In conducting cluster analysis, there are common issues often encountered such as the 'curse of dimensionality' in high-dimensional spaces. This issue can be addressed by performing dimensionality reduction prior to clustering. Food security vulnerability is a global concern for various countries, including Indonesia. According to information from the Food Security Agency, the level of food security in Indonesia in 2022 was at 60.2, which is below the global average of 62.2. In this study, cluster analysis was conducted to map provinces in Indonesia based on 17 food security indicators using the DBSCAN (Density-based Spatial Clustering of Application with Noise) algorithm. Subsequently, a comparison of Silhouette coefficients was made on data resulting from PCA (Principal Component Analysis) reduction and data resulting from t-SNE (t-distributed Stochastic Neighbor Embedding) reduction to obtain the best cluster results. There were 9 clusters formed without any noise formed by data resulting from t-SNE reduction with Silhouette coefficient of 0.7761 that quite strong.
Item Type: | Thesis (Sarjana) |
---|---|
Identification Number: | 0524090169 |
Uncontrolled Keywords: | Clustering, DBSCAN, Ketahanan Pangan, PCA, t-SNE |
Divisions: | Fakultas Matematika dan Ilmu Pengetahuan Alam > Statistika |
Depositing User: | Unnamed user with username nova |
Date Deposited: | 31 Jul 2024 05:01 |
Last Modified: | 31 Jul 2024 05:01 |
URI: | http://repository.ub.ac.id/id/eprint/222208 |
![]() |
Text (DALAM MASA EMBARGO)
Salma Fitri Nugroho.pdf Restricted to Registered users only Download (4MB) |
Actions (login required)
![]() |
View Item |