A machine learning approach to meteor classification

doi:10.1016/j.icarus.2026.117128

A machine learning approach to meteor classification

2026-04-27

0

OA

AI

S

Samantha Hemmelgarn *

DOI：10.1016/j.icarus.2026.117128

Original

Save

View PDF

Abstract

En 中文

• Modern meteor datasets contain observed parameters under-utilized in traditional classification schemes. • Machine learning techniques provide novel methods with which to analyze and interpret meteor datasets. • A framework utilizing Factor Analysis for dimensionality reduction and a Gaussian Mixture Model for clustering recovers expected trends from traditional models. • Factor Analysis reveals that a meteor’s activation threshold is the dominant latent variable governing its inferred material strength. • We present a structural classification scheme, Hclass, that describes the hardness of an observed meteor event.

Keywords:

Asteroids, composition

Comets, composition

Meteors

A machine learning approach to meteor classification

Abstract

AI Summary

Journal

Organization