Machine learning classification of Kuiper belt populations

Abstract

<jats:title>Abstract</jats:title> <jats:p>In the outer solar system, the Kuiper Belt contains dynamical sub-populations sculpted by a combination of planet formation and migration and gravitational perturbations from the present-day giant planet configuration. The subdivision of observed Kuiper Belt objects (KBOs) into different dynamical classes is based on their current orbital evolution in numerical integrations of their orbits. Here we demonstrate that machine learning algorithms are a promising tool for reducing both the computational time and human effort required for this classification. Using a Gradient Boosting Classifier, a type of machine learning regression tree classifier trained on features derived from short numerical simulations, we sort observed KBOs into four broad, dynamically distinct populations–classical, resonant, detached, and scattering–with a &amp;gt;97 per cent accuracy for the testing set of 542 securely classified KBOs. Over 80 per cent of these objects have a &amp;gt;3σ probability of class membership, indicating that the machine learning method is classifying based on the fundamental dynamical features of each population. We also demonstrate how, by using computational savings over traditional methods, we can quickly derive a distribution of class membership by examining an ensemble of object clones drawn from the observational errors. We find two major reasons for misclassification: inherent ambiguity in the orbit of the object–for instance, an object that is on the edge of resonance–and a lack of representative examples in the training set. This work provides a promising avenue to explore for fast and accurate classification of the thousands of new KBOs expected to be found by surveys in the coming decade.</jats:p>

A von Mises–Fisher distribution for the orbital poles of the plutinos

“…To see whether our Plutino sample would change when accounting for orbital uncertainties, we used a separate resonant object identification pipeline that began with the same 690-object semimajor axis limits as in Section 4. Our procedure for securely identifying Plutinos does not, like the laborious standard method described in Gladman et al (2008) and used in Volk & Malhotra (2017) and Smullen & Volk (2020), have the advantage of returning to the observations and fitting new orbits from first principles, but it is more portable and more highly automated. We begin by building off the work of Smullen & Volk (2020).…”

“…Our procedure for securely identifying Plutinos does not, like the laborious standard method described in Gladman et al (2008) and used in Volk & Malhotra (2017) and Smullen & Volk (2020), have the advantage of returning to the observations and fitting new orbits from first principles, but it is more portable and more highly automated. We begin by building off the work of Smullen & Volk (2020).…”

“…Full details of the settings used for the machine learning algorithm, and a full explanation and justification of the 55 recorded features, are given in their paper. To allow other researchers to use their gradient boosting classifier to classify KBOs from their barycentric elements without repeating the entire process of selecting and justifying the features of the orbit to record and the settings to use for the algorithm, Smullen & Volk (2020) posted user-friendly sample code and training data to GitHub.…”

Free Inclinations for Transneptunian Objects in the Main Kuiper Belt

“…A human operator then decided to tag it as scattering (a alters more than 1 au), resonant (ϕ ceases to circulate at any moment in the 10 Myr integration), or classical (for a non-scattering and non-resonant particle). Although recent papers have described TNO classification using automatic pipelines (Khain et al 2020) or machine learning algorithms (Smullen & Volk 2020), we decided to do the job manually as the sample was not too large and this remains the most accurate method. Our criterion for resonant objects is quite loose; this is motivated by the fact that even a brief interaction with a meanmotion resonance can significantly alter I free (see Section 2.3).…”

Peculiar orbits and asymmetries in extreme trans-Neptunian space

“…Our approach is essentially different from the one discussed by Smullen & Volk (2020). Known TNOs are classified as classical, resonant, scattered and detached objects (see e.g.…”