ÆGIS is a machine learning pipeline for creating optimized regression models (i.e., emulators) that can predict the outcome of collisions between planets. Each model predicts a specific post-impact property of the collision.
The data comes from 10,700 smoothed-particle hydrodynamics (SPH) simulations of pairwise collisions between rotating, differentiated planets. The collision simulations and associated machine learning pipeline are from Timpe et al. 2020 (in review). The simulation data is publicly available on the Dryad open research repository.