A machine learning method for predicting telescope cycle time applied to the Cerro Murphy Observatory
2024

Predicting Telescope Cycle Time with Machine Learning

Sample size: 7339 publication 10 minutes Evidence: high

Author Information

Author(s): Kicia Mirosław, Kałuszyński Mikołaj, Górski Marek, Chini Rolf, Pietrzyński Grzegorz

Primary Institution: Nicolaus Copernicus Astronomical Center, Polish Academy of Sciences

Hypothesis

Can machine learning methods accurately predict telescope cycle time for efficient observation scheduling?

Conclusion

The machine learning model successfully predicts telescope cycle time, achieving an error of no more than 12 minutes for 84.2% of observation nights.

Supporting Evidence

  • The model was trained on 7339 observations, achieving a high R-squared value of 0.9980.
  • 84.2% of predictions had an error of less than 2%, equating to a maximum of 12 minutes for a 10-hour observation night.
  • The method is designed to work with various telescopes and instruments, enhancing its applicability.

Takeaway

This study created a computer program that helps astronomers know how long it will take to use a telescope, making it easier to plan their observations.

Methodology

The study used a polynomial multiple regression model trained on real telescope data to predict operation time for observations.

Limitations

The model's accuracy may vary with changes in telescope and instrument parameters, and historical data may not always be reliable.

Digital Object Identifier (DOI)

10.1007/s10686-024-09970-8

Want to read the original?

Access the complete publication on the publisher's website

View Original Publication