Using machine learning to hierarchically model the first APOKASC catalogue of Kepler dwarf and subgiant stars (Serenelli et al. 2017). A new method which encodes the population distribution of helium and mixing-length parameter in the solar neighbourhood.
This repository is the working notebook concerning the hierarchical modelling of low-mass dwarfs in the solar neighbourhood using the machine learning of stellar models. The repository contains Jupyter notebooks and python scripts written for the project.
Requires Python >= 3.5 and a number of packages found in hierarchically-modelling-dwarfs/requirements.txt. An unpublished package called interstellar is also used for the work in hierarchically-modelling-dwarfs/training and hierarchically-modelling-dwarfs/modelling. Whilst this package is not publicly available, please contact me for details if you wish to run this code. We aim to publish the package in the future if it would be useful to the wider community.
To download the repository, run the following command in a terminal,
git clone https://github.com/alexlyttle/hierarchically-modelling-dwarfs.git
Much of the work is written in Jupyter notebooks and may require light editing (e.g. custom paths to data not contained in the repo).
If you would like to details on accessing some external data, please contact me. For example, the training dataset for the neural network would be too large to contain in this repository (please see the paper for details on reproducing the training data).
This repo is not intended to be regularly maintained but contributions are welcome. I also encourage you to raise an Issue if you have a question or concern about the work.
Lyttle et al. (2021; preprint)
The preferred way to contact me with issues directly related to the code in this repository is by submitting an issue. Otherwise, see the contact information below:
Name: Alex Lyttle
Email: [email protected]
GitHub: alexlyttle
Twitter: @_alexlyttle
This work is a part of a project that has received funding from the European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation programme (CartographY; grant agreement ID 804752).