Skip to main content
SearchLoginLogin or Signup

DiSCo Spectral Clasiffier: An Innovative Tool to Understand Ices in the Solar System

Presentation #301.05 in the session TNOs and Centaurs as Viewed by JWST (Oral Presentation)

Published onOct 23, 2023
DiSCo Spectral Clasiffier: An Innovative Tool to Understand Ices in the Solar System

We present the spectral classifier for JWST/ NIRSpec spectra. DiSCo-TNOs is the JWST-GO1 large Program designed to study the surface compositional of trans-Neptunian objects. Recently, Pinilla-Alonso et al. 2023 (in preparation) show that the spectra of TNOs are distributed in four spectral and compositional groups. These groups, named after the shape of the spectra around 3 microns, are indicative of evident differences in surface composition that Pinilla-Alonso et al. interpret in respect of the snow lines in the primordial Solar System.

As a tool to study the future spectra of trans-Neptunian objects, centaurs, and other icy small bodies, we are developing a spectra classifier using various machine learning-based classification techniques such as Support Vector Machines (SVM) or Density-Based Spatial Clustering of Applications with Noise (DBSCAN), and a novel approach for adapting a regression-based machine learning algorithm called Multivariate Adaptive Regression Splines (MARS) for classification purposes.

In this work, we present the early results of this adapted technique, specifically designed for the identification of new clusters and the integration of reinforcement learning into the tool. In the future, we will offer this tool to the community as a way to enable a broad study of the distribution of ice in the outer Solar System.

Comments
0
comment
No comments here