Classification of orbital anomalies using machine learning models

Abstract Orbital anomalies, such as unexpected changes in a satellite’s orbit, can significantly impact satellite operations and threaten human and financial resources. Detecting these anomalies is a crucial challenge in data analysis within the spatial domain. During satellite operations, a large amount of data is generated, These data allow us to assess the satellite’s state at a specific moment. This paper introduces a study aimed at developing a diagnostic algorithm using machine learning techniques to classify orbital anomalies. The proposed algorithm will be able to classify these anomalies based on their origins using various supervised learning models. The performance of the algorithms used was measured by comparing their accuracy and using the confusion matrix over five distinct classes of orbital anomalies.