Earl Bellinger's Ph.D. repository on forward and inverse problems in asteroseismology.
If any of these programs are useful to you, please consider citing one or more of the following:
-
Bellinger, E. P., Angelou, G. C., Hekker, S., Basu, S., Ball, W., Guggenberger, E. (2016). Fundamental Parameters of Main-Sequence Stars in an Instant with Machine Learning. The Astrophysical Journal, 830 (1), 20.
-
Angelou, G. C., Bellinger, E. P., Hekker, S., Basu, S. (2017). On the Statistical Properties of the Lower Main Sequence. The Astrophysical Journal, 839 (2), 116.
-
Bellinger, E. P., Angelou, G., Hekker, S., Basu, S., Ball, W., Guggenberger, E. (2017). Fundamental Parameters in an Instant with Machine Learning: Application to Kepler LEGACY Targets. Seismology of the Sun and Distant Stars, European Physical Journal Web of Conferences.
-
Bellinger, E. P., Basu, S., Hekker, S., Ball, W. (2017). Model-independent Measurement of Internal Stellar Structure in 16 Cygni A and B. The Astrophysical Journal, 851 (2), 80.
See Releases for the versions of this repository corresponding to those papers.
forward/
-
mesa_template/
-- directory containing default instructions for a MESA evolutionary track (copied automatically bydispatch
) -
python3 sobol_dispatcher.py
-- generate initial conditions varied in a quasi-random fashion, calls the following files: *./dispatch.sh
-- shell script for generating a parameterized MESA evolutionary track (e.g. -M 1 for a 1 solar mass model) *Rscript discontinuity.R
-- detect discontinuities in the simulated evolution *Rscript summarize.R
-- summarize an evolutionary track into a matrix -
Rscript collate.R
-- collect nearly-evenly-spaced points from each summarized simulation into one big data filesimulations.dat
; this facilitates the inverse problem -
./amp.sh
-- Runsobol_dispatcher.py
with settings that facilitate comparison with the asteroseismic modeling portal (AMP)
analyze/
-
Rscript diffusion.R
-- plots the initial and final simulation metallicities as a function of mass and diffusion -
Rscript inputs.R
-- creates a diagram showing the initial conditions of the grid based off of../forward/initial_conditions.dat
which is generated bysobol_dispatcher.py
regression/
-
Rscript tagesstern.R
-- degrade BiSON solar frequencies to the level of what is observable from the 16 Cyg stars for the sake of fair evaluation & comparison -
Rscript hare_compile.R
-- turn the Hare & Hound data into a format I can parse -
Rscript perturb.R
-- make Monte-Carlo perturbations of solar, Tagesstern, 16 Cyg, kages, and hares data to account for uncertainties in observed data -
python3 subsetter.py
-- determine the number of evolutionary tracks, models per evolutionary track, and trees in the forest that are needed forlearn.py
to work well *Rscript forest_evaluate.R
-- visualize the output ofsubsetter.py
-
python3 learn.py
-- learn what relates observable data to model properties from../forward/simulations.dat
and predict the properties of the stars inperturb/
*Rscript importances.R
-- plots the feature importances of the random forests obtained inlearn.py
*Rscript cyg.R
-- plots the predicted quantities of 16 Cyg fromlearn.py
against literature values *Rscript us-vs-them.R
-- plots the predicted quantities of the KAGES stars and the Hare-and-Hound exercise against the literature values; also creates the diffusion plot for the KAGES stars -
Rscript legacy.R
-- plots the cumulative distribution functions for estimate uncertainties for the LEGACY targets
inversion/
- Coming soon...
scripts/
maybe_sub.sh
-- shell script for submitting jobs to the condor queuing systemfgong2freqs.sh
-- shell script for redistributing a MESA model mesh and calculating adiabatic pulsation frequencies via ADIPLSseismology.R
-- R script for making seismological calculations from a frequencies data fileutils.R
-- R utility script for plotting, constants, etcsobol_lib.py
-- python library for generating Sobol (quasi-random) numberskerexact.sh
-- generates kernel functions from stellar models
misc/
jcd-kasc/
Rscript CD_diagram.R
-- plot an asteroseismic H-R diagram from a grid of MESA/GYRE models and overplot LEGACY data points on it
python3 plot_sph_harm.py
-- make spherical harmonics plots to visualize the pulsation frequencies of solar-like oscillatorspython3 plot_grids.py
-- make plots of linear, random, and quasi-random (Sobol) grids to justify the use of the latterpython3 plot_classification.py
-- make plots of linear and non-linear (in this case, XOR) classification problems to illustrate the limitations and usefulness of basic and advanced ML routinesRscript 16CygB.R
-- make an annotated power spectrum of 16 Cyg Bmatlab animate_sph_harm.m
-- create animations of spherical harmonicsRscript interp_vs_reg.R
-- plot the difference between linear interpolation and regressionRscript plot_nearly-even-spacing.R
-- show the result of the linear transport problem on finding nearly-evenly spaced points