View Item 
      •   IPB Repository
      • Dissertations and Theses
      • Undergraduate Theses
      • UT - Faculty of Forestry and Environment
      • UT - Forest Management
      • View Item
      •   IPB Repository
      • Dissertations and Theses
      • Undergraduate Theses
      • UT - Faculty of Forestry and Environment
      • UT - Forest Management
      • View Item
      JavaScript is disabled for your browser. Some features of this site may not work without it.

      Algoritma Pohon Keputusan Pembelajar Mesin (Decision Tree of Machine Learning) dalam Mendeteksi Hutan Mangrove Menggunakan Citra SPOT 7

      No Thumbnail [100%x80]
      View/Open
      Cover (985.5Kb)
      Fullteks (6.232Mb)
      Lampiran (794.0Kb)
      Date
      2023-09-26
      Author
      Prabowo, Haidar Nashir Ramadhani
      Jaya, I Nengah Surati
      Metadata
      Show full item record
      Abstract
      Tulisan ini menerangkan tentang pembangunan algoritma klasifikasi non-parametrik berbasis pohon keputusan dari pembelajar mesin dalam mengidentifikasi tutupan pada ekosistem mangrove menggunakan data penginderaan jauh dan data geospasial. Tujuan dari penelitian ini adalah untuk menemukan peubah (atribut) serta parameter algoritma pohon keputusan dalam mendeteksi dan mengidentifikasi tutupan hutan mangrove berbasis citra resolusi tinggi SPOT 7. Kajian ini menguji peubah-peubah geo-sosio-biofisik dan peubah spektral yang diturunkan dari citra SPOT 7. Penelitian ini menemukan bahwa peubah paling berpengaruh adalah elevasi dan NDVI. Penelitian ini juga menemukan parameter dari algoritma pohon keputusan terbaik adalah gain ratio tanpa kombinasi pemangkasan (pruning) dan pra-pangkas (prepruning), kedalaman pohon 31, alternatif prepruning 30, sampling data split dengan otomatis dan ukuran daun 61. Kombinasi tersebut menghasilkan overall accuracy (OA) sebesar 93.7% dan kappa accuracy (KA) 93.2%. Penelitian ini menyimpulkan bahwa peubah geo-sosio-biofisik lebih handal perannya dalam mengidentifikasi kelas-kelas tutupan lahan yang dibangun dibandingkan peubah indeks vegetasi.
       
      This paper describes the development of a decision tree machine learning, a non-parametric classification algorithm for identifying mangrove forest using remote sensing and geospatial data. The study objective is to find the decision tree algorithm's variables (attributes) and parameters in detecting mangrove forest cover based on SPOT 7 high-resolution imagery. This study examined geo-socio-biophysicalvariables and spectral variables derived from SPOT 7 imagery. This study also found that the best decision tree algorithm parameters were the gain ratio without a combination of pruning and pre-pruning, tree depth of 31, alternative prepruning 30, split data sampling with automatic, and leaf size of 41. The algorithm provided an overall accuracy (OA) of 93.7% and a Kappa accuracy (KA) of 93.2%. This study concludes that the geo-socio-biophysicalvariable is more reliable than vegetation index variables in identifying built-up land cover classes.
       
      URI
      http://repository.ipb.ac.id/handle/123456789/125467
      Collections
      • UT - Forest Management [3097]

      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
        

       

      NoThumbnail