SPOCK 2.0: Update to the FeatureClassifier in the Stability of Planetary Orbital Configurations Klassifier

TL;DR

Predicting long-term planetary system stability with direct N-body integrations is computationally expensive. SPOCK 2.0 retrains the FeatureClassifier on a cleaned dataset, computes a system-specific fastest secular timescale $T_{sec}$ from analytic approximations, integrates short data up to $T_{sec}$, and includes $T_{sec}$ as a feature, with updated REBOUND MEGNO handling. Key findings include an AUC of $0.950$ when using $T_{sec}$ both as the integration horizon and as a feature, a cleaned dataset containing 102,497 resonant and 24,941 random configurations, and mitigated MEGNO overflow through REBOUND updates. This work yields faster, more accurate stability predictions tailored to a system's secular dynamics and provides a transparent, reproducible data and API update for the community.

Abstract

The Stability of Planetary Orbital Configurations Klassifier (SPOCK) package collects machine learning models for predicting the stability and collisional evolution of compact planetary systems. In this paper we explore improvements to SPOCK's binary stability classifier (FeatureClassifier), which predicts orbital stability by collecting data over a short N-body integration of a system. We find that by using a system-specific timescale (rather than a fixed $10^4$ orbits) for the integration, and by using this timescale as an additional feature, we modestly improve the model's AUC metric from 0.943 to 0.950 (AUC=1 for a perfect model). We additionally discovered that $\approx 10\%$ of N-body integrations in SPOCK's original training dataset were duplicated by accident, and that $<1\%$ were misclassified as stable when they in fact led to ejections. We provide a cleaned dataset of 100,000+ unique integrations, release a newly trained stability classification model, and make minor updates to the API.