Image Details

Choose export citation format:

Machine Learning for Radial Velocity Analysis. I. Vision Transformers as a Robust Alternative for Detecting Planetary Candidates

  • Authors: Anoop Gavankar, Tanish Mittal, Joe P. Ninan, Shravan Hanasoge

Anoop Gavankar et al 2026 The Astronomical Journal 171 .

  • Provider: AAS Journals

Caption: Figure 19.

For the ordered data set V3, accuracy and precision (as described previously in Figure 13) offer complementary insights into the model’s performance in predicting orbital periods. Accuracy declines near values corresponding to the solar rotation rate, suggesting that the rotational signal retains some ambiguity despite fine-tuning, resulting in frequent misclassifications around this period. In contrast, precision increases monotonically with the orbital period. This upward trend reflects a systematic bias where misclassifications are skewed toward lower period values. Consequently, high-period predictions are less likely to be incorrectly assigned to shorter periods, leading to improved precision at longer orbital periods. This pattern suggests that while the model struggles to differentiate signals near the stellar rotation period, it demonstrates greater confidence and reliability in its high-period classifications.

Other Images in This Article

Show More

Copyright and Terms & Conditions

Additional terms of reuse