Simplifying asteroseismic analysis of solar-like oscillators: An application of principal component analysis for dimensionality reduction

M. B. Nielsen; G. R. Davies; W. J. Chaplin; W. H Ball; J. M. J. Ong; E. Hatt; B. P. Jones; M. Logue

doi:10.1051/0004-6361/202346086

Simplifying asteroseismic analysis of solar-like oscillators: An application of principal component analysis for dimensionality reduction

M. B. Nielsen, G. R. Davies, W. J. Chaplin, W. H Ball, J. M. J. Ong, E. Hatt, B. P. Jones, M. Logue

Physics and Astronomy

Research output: Contribution to journal › Article › peer-review

40 Downloads (Pure)

Abstract

Context. The asteroseismic analysis of stellar power density spectra is often computationally expensive. The models used in the analysis may require several dozen parameters to accurately describe features in the spectra caused by the oscillation modes and surface granulation. Many of these parameters are often highly correlated, making the parameter space difficult to quickly and accurately sample. They are, however, all dependent on a much smaller set of parameters, namely the fundamental stellar properties.

Aims. We aim to leverage this to develop a method for simplifying the process of sampling the model parameter space for the asteroseismic analysis of solar-like oscillators, with an emphasis on mode identification.

Methods. Using a large set of previous observations, we applied principal component analysis to the sample covariance matrix to select a new basis on which to sample the model parameters. Selecting the subset of basis vectors that explains the majority of the sample variance, we then redefined the model parameter prior probability density distributions in terms of a smaller set of latent parameters.

Results. We are able to reduce the dimensionality of the sampled parameter space by a factor of two to three. The number of latent parameters needed to accurately model the stellar oscillation spectra cannot be determined exactly but is likely only between four and six. Using two latent parameters, the method is able to produce models that describe the bulk features of the oscillation spectrum, while including more latent parameters allows for a frequency precision better than ≈ 10% of the small frequency separation for a given target.

Conclusions. We find that sampling a lower-rank latent parameter space still allows for accurate mode identification and parameter estimation on solar-like oscillators over a wide range of evolutionary stages. This allows for the potential to increase the complexity of spectrum models without a corresponding increase in computational expense.

Original language	English
Article number	A117
Number of pages	11
Journal	Astronomy and Astrophysics
Volume	676
DOIs	https://doi.org/10.1051/0004-6361/202346086
Publication status	Published - 22 Aug 2023

Bibliographical note

Acknowledgements:
Thanks to Chris Moore for the useful and informative chats. M.B.N. acknowledges support from the UK Space Agency. G.R.D. and W.J.C. acknowledge the support of the UK Science and Technology Facilities Council (STFC). J.O. acknowledges support from NASA through the NASA Hubble Fellowship grant HST-HF2-51517.001 awarded by the Space Telescope Science Institute, which is operated by the Association of Universities for Research in Astronomy, Incorporated, under NASA contract NAS5-26555. This paper has received funding from the European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation programme (CartographY GA. 804752). The authors acknowledge use of the Blue-BEAR HPC service at the University of Birmingham. This paper includes data collected by the Kepler mission and obtained from the MAST data archive at the Space Telescope Science Institute (STScI). Funding for the Kepler mission is provided by the NASA Science Mission Directorate. STScI is operated by the Association of Universities for Research in Astronomy, Inc., under NASA contract NAS5-26555. This paper includes data collected by the TESS mission. Funding for the TESS mission is provided by the NASA’s Science Mission Directorate. This work has made use of data from the European Space Agency (ESA) mission Gaia (https://www.cosmos.esa.int/gaia), processed by the Gaia Data Processing and Analysis Consortium (DPAC, https://www.cosmos.esa.int/web/gaia/dpac/consortium). Funding for the DPAC has been provided by national institutions, in particular the institutions participating in the Gaia Multilateral Agreement. This publication makes use of data products from the Two Micron All Sky Survey, which is a joint project of the University of Massachusetts and the Infrared Processing and Analysis Center/California Institute of Technology, funded by the National Aeronautics and Space Administration and the National Science Foundation.

Keywords

Asteroseismology
Stars: oscillations
Methods: data analysis
Methods: statistical

Access to Document

10.1051/0004-6361/202346086Licence: Creative Commons: Attribution (CC BY)

NielsenM2023SimplifyingFinal published version, 4.91 MBLicence: Creative Commons: Attribution (CC BY)

H2020_ERC_CARTOGRAPHY
Davies, G.
European Commission
1/04/19 → 31/03/25
Project: EU

Cite this

@article{5f65f38af68a4dd0af3e28f930d94bed,

title = "Simplifying asteroseismic analysis of solar-like oscillators: An application of principal component analysis for dimensionality reduction",

abstract = "Context. The asteroseismic analysis of stellar power density spectra is often computationally expensive. The models used in the analysis may require several dozen parameters to accurately describe features in the spectra caused by the oscillation modes and surface granulation. Many of these parameters are often highly correlated, making the parameter space difficult to quickly and accurately sample. They are, however, all dependent on a much smaller set of parameters, namely the fundamental stellar properties.Aims. We aim to leverage this to develop a method for simplifying the process of sampling the model parameter space for the asteroseismic analysis of solar-like oscillators, with an emphasis on mode identification.Methods. Using a large set of previous observations, we applied principal component analysis to the sample covariance matrix to select a new basis on which to sample the model parameters. Selecting the subset of basis vectors that explains the majority of the sample variance, we then redefined the model parameter prior probability density distributions in terms of a smaller set of latent parameters.Results. We are able to reduce the dimensionality of the sampled parameter space by a factor of two to three. The number of latent parameters needed to accurately model the stellar oscillation spectra cannot be determined exactly but is likely only between four and six. Using two latent parameters, the method is able to produce models that describe the bulk features of the oscillation spectrum, while including more latent parameters allows for a frequency precision better than ≈ 10% of the small frequency separation for a given target.Conclusions. We find that sampling a lower-rank latent parameter space still allows for accurate mode identification and parameter estimation on solar-like oscillators over a wide range of evolutionary stages. This allows for the potential to increase the complexity of spectrum models without a corresponding increase in computational expense.",

keywords = "Asteroseismology, Stars: oscillations, Methods: data analysis, Methods: statistical",

author = "Nielsen, {M. B.} and Davies, {G. R.} and Chaplin, {W. J.} and Ball, {W. H} and Ong, {J. M. J.} and E. Hatt and Jones, {B. P.} and M. Logue",

note = "Acknowledgements: Thanks to Chris Moore for the useful and informative chats. M.B.N. acknowledges support from the UK Space Agency. G.R.D. and W.J.C. acknowledge the support of the UK Science and Technology Facilities Council (STFC). J.O. acknowledges support from NASA through the NASA Hubble Fellowship grant HST-HF2-51517.001 awarded by the Space Telescope Science Institute, which is operated by the Association of Universities for Research in Astronomy, Incorporated, under NASA contract NAS5-26555. This paper has received funding from the European Research Council (ERC) under the European Union{\textquoteright}s Horizon 2020 research and innovation programme (CartographY GA. 804752). The authors acknowledge use of the Blue-BEAR HPC service at the University of Birmingham. This paper includes data collected by the Kepler mission and obtained from the MAST data archive at the Space Telescope Science Institute (STScI). Funding for the Kepler mission is provided by the NASA Science Mission Directorate. STScI is operated by the Association of Universities for Research in Astronomy, Inc., under NASA contract NAS5-26555. This paper includes data collected by the TESS mission. Funding for the TESS mission is provided by the NASA{\textquoteright}s Science Mission Directorate. This work has made use of data from the European Space Agency (ESA) mission Gaia (https://www.cosmos.esa.int/gaia), processed by the Gaia Data Processing and Analysis Consortium (DPAC, https://www.cosmos.esa.int/web/gaia/dpac/consortium). Funding for the DPAC has been provided by national institutions, in particular the institutions participating in the Gaia Multilateral Agreement. This publication makes use of data products from the Two Micron All Sky Survey, which is a joint project of the University of Massachusetts and the Infrared Processing and Analysis Center/California Institute of Technology, funded by the National Aeronautics and Space Administration and the National Science Foundation.",

year = "2023",

month = aug,

day = "22",

doi = "10.1051/0004-6361/202346086",

language = "English",

volume = "676",

journal = "Astronomy and Astrophysics",

issn = "0004-6361",

publisher = "EDP Sciences",

}

TY - JOUR

T1 - Simplifying asteroseismic analysis of solar-like oscillators

T2 - An application of principal component analysis for dimensionality reduction

AU - Nielsen, M. B.

AU - Davies, G. R.

AU - Chaplin, W. J.

AU - Ball, W. H

AU - Ong, J. M. J.

AU - Hatt, E.

AU - Jones, B. P.

AU - Logue, M.

N1 - Acknowledgements: Thanks to Chris Moore for the useful and informative chats. M.B.N. acknowledges support from the UK Space Agency. G.R.D. and W.J.C. acknowledge the support of the UK Science and Technology Facilities Council (STFC). J.O. acknowledges support from NASA through the NASA Hubble Fellowship grant HST-HF2-51517.001 awarded by the Space Telescope Science Institute, which is operated by the Association of Universities for Research in Astronomy, Incorporated, under NASA contract NAS5-26555. This paper has received funding from the European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation programme (CartographY GA. 804752). The authors acknowledge use of the Blue-BEAR HPC service at the University of Birmingham. This paper includes data collected by the Kepler mission and obtained from the MAST data archive at the Space Telescope Science Institute (STScI). Funding for the Kepler mission is provided by the NASA Science Mission Directorate. STScI is operated by the Association of Universities for Research in Astronomy, Inc., under NASA contract NAS5-26555. This paper includes data collected by the TESS mission. Funding for the TESS mission is provided by the NASA’s Science Mission Directorate. This work has made use of data from the European Space Agency (ESA) mission Gaia (https://www.cosmos.esa.int/gaia), processed by the Gaia Data Processing and Analysis Consortium (DPAC, https://www.cosmos.esa.int/web/gaia/dpac/consortium). Funding for the DPAC has been provided by national institutions, in particular the institutions participating in the Gaia Multilateral Agreement. This publication makes use of data products from the Two Micron All Sky Survey, which is a joint project of the University of Massachusetts and the Infrared Processing and Analysis Center/California Institute of Technology, funded by the National Aeronautics and Space Administration and the National Science Foundation.

PY - 2023/8/22

Y1 - 2023/8/22

N2 - Context. The asteroseismic analysis of stellar power density spectra is often computationally expensive. The models used in the analysis may require several dozen parameters to accurately describe features in the spectra caused by the oscillation modes and surface granulation. Many of these parameters are often highly correlated, making the parameter space difficult to quickly and accurately sample. They are, however, all dependent on a much smaller set of parameters, namely the fundamental stellar properties.Aims. We aim to leverage this to develop a method for simplifying the process of sampling the model parameter space for the asteroseismic analysis of solar-like oscillators, with an emphasis on mode identification.Methods. Using a large set of previous observations, we applied principal component analysis to the sample covariance matrix to select a new basis on which to sample the model parameters. Selecting the subset of basis vectors that explains the majority of the sample variance, we then redefined the model parameter prior probability density distributions in terms of a smaller set of latent parameters.Results. We are able to reduce the dimensionality of the sampled parameter space by a factor of two to three. The number of latent parameters needed to accurately model the stellar oscillation spectra cannot be determined exactly but is likely only between four and six. Using two latent parameters, the method is able to produce models that describe the bulk features of the oscillation spectrum, while including more latent parameters allows for a frequency precision better than ≈ 10% of the small frequency separation for a given target.Conclusions. We find that sampling a lower-rank latent parameter space still allows for accurate mode identification and parameter estimation on solar-like oscillators over a wide range of evolutionary stages. This allows for the potential to increase the complexity of spectrum models without a corresponding increase in computational expense.

AB - Context. The asteroseismic analysis of stellar power density spectra is often computationally expensive. The models used in the analysis may require several dozen parameters to accurately describe features in the spectra caused by the oscillation modes and surface granulation. Many of these parameters are often highly correlated, making the parameter space difficult to quickly and accurately sample. They are, however, all dependent on a much smaller set of parameters, namely the fundamental stellar properties.Aims. We aim to leverage this to develop a method for simplifying the process of sampling the model parameter space for the asteroseismic analysis of solar-like oscillators, with an emphasis on mode identification.Methods. Using a large set of previous observations, we applied principal component analysis to the sample covariance matrix to select a new basis on which to sample the model parameters. Selecting the subset of basis vectors that explains the majority of the sample variance, we then redefined the model parameter prior probability density distributions in terms of a smaller set of latent parameters.Results. We are able to reduce the dimensionality of the sampled parameter space by a factor of two to three. The number of latent parameters needed to accurately model the stellar oscillation spectra cannot be determined exactly but is likely only between four and six. Using two latent parameters, the method is able to produce models that describe the bulk features of the oscillation spectrum, while including more latent parameters allows for a frequency precision better than ≈ 10% of the small frequency separation for a given target.Conclusions. We find that sampling a lower-rank latent parameter space still allows for accurate mode identification and parameter estimation on solar-like oscillators over a wide range of evolutionary stages. This allows for the potential to increase the complexity of spectrum models without a corresponding increase in computational expense.

KW - Asteroseismology

KW - Stars: oscillations

KW - Methods: data analysis

KW - Methods: statistical

U2 - 10.1051/0004-6361/202346086

DO - 10.1051/0004-6361/202346086

M3 - Article

SN - 0004-6361

VL - 676

JO - Astronomy and Astrophysics

JF - Astronomy and Astrophysics

M1 - A117

ER -

Simplifying asteroseismic analysis of solar-like oscillators: An application of principal component analysis for dimensionality reduction

Abstract

Bibliographical note

Keywords

Access to Document

Fingerprint

Projects

H2020_ERC_CARTOGRAPHY

Cite this