A machine learning method for predicting telescope cycle time applied to the Cerro Murphy Observatory.

Area of Science:

• Astronomy
• Observational Astrophysics
• Data Science

Background:

• Accurate telescope cycle time estimation is crucial for observational astronomy.
• Existing tools for calculating observation time for multiple telescopes and instruments are limited.
• Optimizing observatory efficiency requires reliable methods for predicting observation durations.

Purpose of the Study:

• To develop and present a novel, automated tool for determining telescope observation time.
• To implement a Machine Learning (ML) based software module for observatory planning.
• To demonstrate the application and effectiveness of the tool at the Cerro Murphy Observatory (OCM) and its general applicability.

Main Methods:

• Utilized a polynomial multiple regression method for time estimation.
• Developed a fully automatic software module requiring no user intervention.
• Detailed the process of building a reliable ML model, including data collection, cleaning, training, and error evaluation.

Main Results:

• The ML model was successfully implemented within observatory software.
• Accuracy analysis using real telescope data confirmed the method's usefulness.
• Prediction errors for 84.2% of nights did not exceed 2%, equating to a maximum 12-minute error in a 10-hour night.

Conclusions:

• The developed ML tool effectively optimizes observation efficiency at astronomical observatories.
• The presented methodology for building and evaluating the ML model is robust and applicable to various telescope-instrument configurations.
• The automated system provides accurate and reliable telescope cycle time estimations, aiding astronomical research planning.