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.

      Perbandingan Performa Long Short-Term Memory dan XGBoost dalam Memprediksi Curah Hujan Harian di Sumatera Selatan

      Thumbnail
      View/Open
      Cover (489.4Kb)
      Fulltext (3.047Mb)
      Lampiran (238.3Kb)
      Date
      2024
      Author
      HADI, OKSI AL
      Masjkur, Mohammad
      Sadik, Kusman
      Metadata
      Show full item record
      Abstract
      Bencana di Indonesia didominasi oleh bencana hidrometeorologi seperti banjir, tanah longsor, dan kekeringan. Salah satu faktor penyebab bencana hidrometeorologi adalah curah hujan. Tercatat telah terjadi 60 kasus bencana hidrometeorologi di Sumatera Selatan yaitu banjir sebanyak 43 kali dan tanah longsor sebanyak 17 kali. Terdapat berbagai algoritme machine learning yang memiliki kompleksitas tinggi dan mampu menghasilkan akurasi terbaik pada banyak kasus seperti long short-term memory dan XGBoost. Penelitian ini bertujuan untuk membandingkan performa algoritme long short-term memory dan XGBoost serta menentukan peubah-peubah yang berpengaruh terhadap curah hujan harian melalui metode shapley additive explanations. Dataset yang digunakan merupakan data harian di Provinsi Sumatera Selatan periode 2011 hingga 2020. Peubah yang digunakan ada enam yaitu curah hujan sebagai peubah respon dan lima peubah penjelas lainnya yaitu temperatur, kelembapan udara, kecepatan angin, tekanan udara, dan titik embun. Dalam membandingkan kedua model tersebut, XGBoost memiliki performa yang lebih unggul dibandingkan dengan LSTM dalam memprediksi curah hujan dengan nilai RMSE dan MAE pada hasil prediksi sebesar 5,21 dan 1,93. Peubah penting yang dihasilkan pada pemodelan XGBoost adalah rataan dua data curah hujan sebelumnya dan curah hujan satu hari sebelumnya.
       
      Disasters in Indonesia are dominated by hydrometeorological disasters such as floods, landslides and droughts. One of the factors causing hydrometeorological disasters is rainfall. There have been 60 cases of hydrometeorological disasters in South Sumatra, namely floods 43 times and landslides 17 times. There are various machine learning algorithms that have high complexity and are able to produce the best accuracy in many cases such as long short-term memory and XGBoost. This study aims to compare the performance of long short-term memory and XGBoost algorithms and determine the variables that affect daily rainfall through the shapley additive explanations method. The dataset used is daily data in South Sumatra Province for the period 2011 to 2020. There are six variables used, namely rainfall as the response variable and five other explanatory variables, namely temperature, air humidity, wind speed, air pressure, and dew point. In comparing the two models, XGBoost has superior performance compared to LSTM in predicting rainfall with RMSE and MAE values in the prediction results of 5,21 and 1,93. The important variables generated in XGBoost modelling are the average of the two previous rainfall data and the rainfall one day before.
       
      URI
      http://repository.ipb.ac.id/handle/123456789/157962
      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