View Item 
      •   IPB Repository
      • Dissertations and Theses
      • Undergraduate Theses
      • UT - Faculty of Mathematics and Natural Sciences
      • UT - Statistics and Data Sciences
      • View Item
      •   IPB Repository
      • Dissertations and Theses
      • Undergraduate Theses
      • UT - Faculty of Mathematics and Natural Sciences
      • UT - Statistics and Data Sciences
      • View Item
      JavaScript is disabled for your browser. Some features of this site may not work without it.

      Penerapan Algoritma DBSCAN untuk Penggerombolan Kota/Kabupaten di Jawa Barat Berdasarkan Indikator Rawan Pangan

      Thumbnail
      View/Open
      Cover (2.731Mb)
      Fullteks (2.726Mb)
      Lampiran (2.731Mb)
      Date
      2022
      Author
      Kevin, Alexander
      Rizki, Akbar
      Sartono, Bagus
      Metadata
      Show full item record
      Abstract
      Terdapat beberapa permasalahan umum ketika melakukan analisis gerombol, diantaranya adalah adanya “curse of dimensionality” di ruang dimensi tinggi dan adanya bentuk gerombol yang tidak beraturan (arbitrary) serta kehadiran outlier. Permasalahan tersebut dapat diatasi dengan melakukan reduksi dimensi terlebih dahulu dan menggunakan analisis gerombol yang berbasis kepadatan. Kerawanan pangan merupakan salah satu permasalahan utama yang menjadi isu global bagi sebagian besar negara. Di Indonesia terdapat FSVA (Food Security and Vulnerability Atlas) yang merupakan peta tematik visualisasi geografis wilayah rentan terhadap kerawanan pangan. Penentuan status rawan pangan suatu daerah pada FSVA diukur dengan analisis yang relatif sederhana yaitu metode pembobotan pada sembilan indikator kerawanan pangan kronis. Pada penelitian kali ini, penggerombolan wilayah yang rentan terhadap kerawanan pangan untuk kota/kabupaten di Jawa Barat dilakukan dengan metode machine learning dan indikator yang digunakan lebih beragam (25 peubah) untuk menghasilkan gerombol yang lebih baik. Penggerombolan dilakukan dengan metode DBSCAN (DensityBased Spatial Clustering of Applications with Noise) pada data yang telah direduksi dimensi terlebih dahulu menggunakan metode t-SNE (t-Distributed Stochastic Neighbor Embedding). Data yang digunakan dalam penelitian ini adalah data Susenas (Survei Sosial Ekonomi Nasional) tahun 2020 untuk provinsi Jawa Barat. Hasil penggerombolan terbaik ditentukan berdasarkan parameter optimal yang dievaluasi menggunakan koefisien Silhouette dan indeks DBCV (Density Based Clustering Validation). Parameter optimal yang menghasilkan koefisien Silhouette dan indeks DBCV tertinggi diperoleh pada perplexity = 1, epsilon = 31, dan minPts = 2. Penggerombolan terbaik menghasilkan lima gerombol dengan tidak ada outlier.
       
      There are several common problems when performing cluster analysis, including the existence of a “curse of dimensionality” in high-dimensional space and the presence of arbitrary cluster shapes, and also the presence of outliers. These problems can be solved by reducing the dimensions first and using density-based cluster analysis. Food insecurity is one of the main problems that become a global issue for most countries. In Indonesia, there is an FSVA (Food Security and Vulnerability Atlas) which is a thematic map visualizing geographical areas vulnerable to food insecurity. Determination of the food insecurity status of an area on the FSVA is measured by relatively simple analysis, namely the weighting method on nine indicators of chronic food insecurity. In this study, clustering of areas that are vulnerable to food insecurity for cities/districts of West Java is carried out using machine learning methods and the indicators used are more diverse (25 variables) to produce better clusters. Clustering is performed using the DBSCAN (Density-Based Spatial Clustering of Applications with Noise) method on data that has been reduced in dimensions first using t-SNE (t-Distributed Stochastic Neighbor Embedding). The data used in this study is the 2020 Susenas (Survei Sosial Ekonomi Nasional) data for the province of West Java. The best clustering result is determined based on the optimal parameters evaluated using the Silhouette coefficient and the DBCV (Density-Based Clustering Validation) index. Optimal parameters that produce the highest Silhouette coefficient and DBCV index are obtained at perplexity = 1, epsilon = 31, and minPts = 2. The best clustering results in five clusters with no outliers.
       
      URI
      http://repository.ipb.ac.id/handle/123456789/113744
      Collections
      • UT - Statistics and Data Sciences [2260]

      Copyright © 2020 Library of IPB University
      All rights reserved
      Contact Us | Send Feedback
      Indonesia DSpace Group 
      IPB University Scientific Repository
      UIN Syarif Hidayatullah Institutional Repository
      Universitas Jember Digital Repository
        

       

      Browse

      All of IPB RepositoryCollectionsBy Issue DateAuthorsTitlesSubjectsThis CollectionBy Issue DateAuthorsTitlesSubjects

      My Account

      Login

      Application

      google store

      Copyright © 2020 Library of IPB University
      All rights reserved
      Contact Us | Send Feedback
      Indonesia DSpace Group 
      IPB University Scientific Repository
      UIN Syarif Hidayatullah Institutional Repository
      Universitas Jember Digital Repository