Image Details

Choose export citation format:

A Machine Learning Approach to Exoplanet Atmospheric Retrieval: Application to Optical Filter Ranking

  • Authors: Patcharawee Munsaket, Supachai Awiphan, Poemwai Chainakun, Eamonn Kerins, Napaporn A-thano

Patcharawee Munsaket et al 2026 The Astronomical Journal 172 .

  • Provider: AAS Journals

Caption: Figure 13.

Comparison of normalization schemes: the commonly used zero-out (mean subtraction) approach versus our adopted z-score standardization. Panels (a) and (b) show the full synthetic transmission spectra, while panels (c) and (d) present the corresponding 10-band photometry. In all cases, the atmospheric parameters are fixed and only the planetary radius varies over Rp = 0.8–1.8RJ. We note that the transit depth is not directly comparable between methods due to different normalizations; therefore, we focus on relative spectral behavior, quantified via the residual variance as illustrated further in Figure 14.

Other Images in This Article

Show More

Copyright and Terms & Conditions

Additional terms of reuse