Unsupervised Method for Correlated Noise Removal for Multi-wavelength Exoplanet Transit Observations

TL;DR

This paper introduces an unsupervised neural-network-based method utilizing clustering to remove correlated noise from multi-wavelength exoplanet transit light curves, enhancing measurement precision in certain cases.

Contribution

A novel adaptive, unsupervised neural network approach combined with clustering techniques for noise removal in exoplanet transit data, improving measurement accuracy.

Findings

Smaller error bars achieved for high-quality data from WASP-19b.

Method did not significantly improve low-quality GJ-1214 data.

Effectiveness depends on data quality and specific observational circumstances.

Abstract

Exoplanetary atmospheric observations require an exquisite precision in the measurement of the relative flux among wavelengths. In this paper, we aim to provide a new adaptive method to treat light curves before fitting transit parameters in order to minimize systematic effects that affect, for instance, ground-based observations of exo-atmospheres. We propose a neural-network-based method that uses a reference built from the data itself with parameters that are chosen in an unsupervised fashion. To improve the performance of proposed method, K-means clustering and Silhouette criteria are used for identifying similar wavelengths in each cluster. We also constrain under which circumstances our method improves the measurement of planetary-to-stellar radius ratio without producing significant systematic offset. We tested our method in high quality data from WASP-19b and low quality data from GJ-1214. We succeed in providing smaller error bars for the former when using JKTEBOP, but GJ-1214 light curve was beyond the capabilities of this method to improve as it was expected from our validation tests.