Akademik Veri Yönetim Sistemi
8th International Management Information Systems Conference (IMISC2021), İstanbul, Türkiye, 20 - 22 Ekim 2021
In this
study, it is aimed to estimate the type of astronomical observation determined
by considering the local characteristics and current observation conditions of
the Eastern Anatolia Observatory. For this purpose, first of all, the data that
created the training set were labeled in line with the expert opinion.
Afterwards, predictions were made on two-class and four-class data using four
different algorithms. As a result of the study, the performance of machine
learning algorithms was measured by estimating the appropriate observation type
for the data of each day. Although it is seen that the algorithms provide over
95% performance for two-class data, the performance of Naive Bayes and K
Nearest Neighbor algorithms has decreased on four-class data. As a result, it is thought that the estimation
studies will contribute to the effective and efficient use of observatories.