Makine öğrenmesi algoritmalarıyla astronomik gözlem kalitesi tahminine yönelik karar destek sistemi geliştirilmesi ve uygulanması


Creative Commons License

Yavuz Ö. Ç., Karaman E., Yeşilyaprak C.

Trends in Business and Economics, cilt.36, sa.3, ss.289-303, 2022 (Hakemli Dergi)

  • Yayın Türü: Makale / Tam Makale
  • Cilt numarası: 36 Sayı: 3
  • Basım Tarihi: 2022
  • Doi Numarası: 10.5152/tbe.2022.1049957
  • Dergi Adı: Trends in Business and Economics
  • Derginin Tarandığı İndeksler: TR DİZİN (ULAKBİM)
  • Sayfa Sayıları: ss.289-303
  • Atatürk Üniversitesi Adresli: Evet

Özet

Kurulumunun tamamlanmasıyla birlikte araştırmacıların kullanımına sunulması planlanan Doğu Anadolu Gözlemevi (DAG) teleskobunun etkin ve verimli kullanımı önem arz etmektedir. Bu çalışma kapsamında araştırmacılar tarafından sunulan projelerin, gözlemevinin bulunduğu bölgenin yerel özellikleri dikkate alınarak gözlem türüyle eşleştirilmesi, değerlendirilmesi ve en uygun güne atanmasına yönelik karar destek sistemi geliştirilmesi amaçlanmaktadır. Bu amaç doğrultusunda öncelikle Naive Bayes, K En Yakın Komşu, Karar Ağacı ve Rastgele Orman algoritması kullanılarak dört farklı algoritmanın performansları değerlendirilmiş, yeniden örnekleme yöntemleri uygulanmış ve öz niteliklerin sonuca etkisi incelenmiştir. Sonrasında MAUT yönteminden esinlenilerek her bir proje için yarar fonksiyonu formülünü barındıran fayda değerlerinin hesaplanmasına dayalı karar destek modeli geliştirilmiştir. Fayda değerleri projeler için başarı puanını temsil etmektedir. Projeler, gözlem türüne göre sınıflandırılarak başarı puanına göre büyükten küçüğe sıralanmıştır. Sonrasında önceden tahmin edilen gözlem türleri doğrultusunda projeler önceliklendirilerek ilgili günlere atanmıştır. Geliştirilen karar destek modeli ile teleskobun etkin ve verimli kullanımıyla birlikte değerlendirme sürecinin otomatikleştirilmesi amaçlanmaktadır.

Anahtar Kelimeler: Astronomik gözlem kalitesi, Doğu Anadolu Gözlemevi, tahmin, karar destek sistemleri, makine öğrenmesi

JEL Kodları: C53, C67, C80, L86, S44

The effective and efficient use of the Eastern Anatolia Observatory telescope, which is planned to be made available to researchers upon completion of its installation, is important. Within the scope of this study, it is aimed to develop a decision support system for the projects presented by the researchers to be matched with the observation type, evaluated, and assigned to the most appropriate day, taking into account the local characteristics of the region where the observatory is located. For this purpose, first of all, the performances of four different algorithms were evaluated using naive Bayes, K nearest neighbor, decision tree, and random forest algorithms, resampling methods were applied, and the effect of attributes on the result was examined. Then, inspired by the multi-attribute utility theory method, a decision support model based on the calculation of utility valuescontaining the utility function formula was developed for each project. Benefit values represent success points for projects. The projects were classified according to the type of observation and ordered from the largest to the smallest according to the success score. Afterward, projects were prioritized in line with the predicted types of observations and assigned to the relevant days. With the developed decision support model, it is aimed to automate the evaluation process together with the effective and efficient use of the telescope.

Keywords: Astronomival observation quality, decision support systems, Eastern Anatolia Observatory, estimation, machine learning

JEL Codes: C53, C67, C80, L86, S44