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 Algoritme Genetika dalam Penentuan Indikator Krisis Keuangan Hasil Pemodelan Machine Learning

      Thumbnail
      View/Open
      Cover (370.0Kb)
      Fullteks (892.7Kb)
      Lampiran (485.5Kb)
      Date
      2024
      Author
      Mufrodi, Achmad Ismail
      Sartono, Bagus
      Anisa, Rahma
      Metadata
      Show full item record
      Abstract
      Krisis keuangan adalah situasi ketika terdapat tanda-tanda kesulitan dan kerugian keuangan yang memengaruhi sebagian besar sistem keuangan. Besar dampak yang akan diterima suatu negara dipengaruhi oleh seberapa siap sistem keuangan yang ada dalam menghadapi syok atau tekanan yang terjadi. Diperlukan sebuah deteksi dini dalam mempersiapkan jika terjadi tekanan pada sistem keuangan. Pemodelan machine learning menggunakan data historis krisis keuangan dapat dimanfaatkan untuk mengantisipasi kejadian krisis tersebut. Penelitian ini menggunakan data dari macrohistory.net/database yang terdiri dari systemic financial crises sebagai peubah respon dan 18 indikator makroekonomi dari 17 negara selama 151 tahun. Data tersebut kemudian diolah dengan metode machine learning yang berasal dari tiga pendekatan berbeda (regresi logistik, random forest, dan XGBoost). Tingkat kepentingan peubah dalam mendeteksi krisis keuangan dapat diukur dengan menghitung permutation feature importance. Selanjutnya, penerapan algoritme genetika dimanfaatkan untuk mendapatkan solusi optimal untuk urutan tingkat kepentingan peubah berdasarkan ketiga metode tersebut. Optimasi urutan pemeringkatan menunjukkan bahwa peubah X14 (equity total return), X10 (leverage) dan X4 (short-term interest rate) merupakan tiga peubah paling penting berdasarkan algoritme genetika. Hasil pemeringkatan kepentingan peubah tersebut memiliki kesetujuan dengan ketiga model machine learning, sehingga dapat digunakan sebagai peubah paling signifikan untuk menjadi indikator krisis keuangan.
       
      A financial crisis is characterized by signs of financial distress and losses affecting much of the financial system. The magnitude of the impact a country receives is influenced by its financial system's preparedness to face the occurring shock or pressure. Early detection is needed to prepare for potential pressure on the financial system. Machine learning modeling using historical financial crisis data can be utilized to anticipate crisis events. Data from macrohistory.net/database, consisting of systemic financial crises as a response variable and 18 macroeconomic indicators from 17 countries over 151 years, is used in this research. The data is then processed using machine learning methods derived from three approaches (logistic regression, random forest, and XGBoost). The level of importance of variables in detecting a financial crisis is measured by calculating the permutation feature importance. Subsequently, a genetic algorithm is applied to determine the optimal solution for the order of variable importance levels based on the three methods. Optimization of the ranking order showed that the variables X14 (equity total return), X10 (leverage), and X4 (short-term interest rate) were the three most important variables based on the genetic algorithm. These variable’s importance rankings matched the three machine learning models, making them the most critical financial crisis indicators.
       
      URI
      http://repository.ipb.ac.id/handle/123456789/152933
      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