Data analysis recipes: Fitting a model to data
read the original abstract
We go through the many considerations involved in fitting a model to data, using as an example the fit of a straight line to a set of points in a two-dimensional plane. Standard weighted least-squares fitting is only appropriate when there is a dimension along which the data points have negligible uncertainties, and another along which all the uncertainties can be described by Gaussians of known variance; these conditions are rarely met in practice. We consider cases of general, heterogeneous, and arbitrarily covariant two-dimensional uncertainties, and situations in which there are bad data (large outliers), unknown uncertainties, and unknown but expected intrinsic scatter in the linear relationship being fit. Above all we emphasize the importance of having a "generative model" for the data, even an approximate one. Once there is a generative model, the subsequent fitting is non-arbitrary because the model permits direct computation of the likelihood of the parameters or the posterior probability distribution. Construction of a posterior probability distribution is indispensible if there are "nuisance parameters" to marginalize away.
This paper has not been read by Pith yet.
Forward citations
Cited by 24 Pith papers
-
Observational signatures of thermonuclear electron-capture supernovae -- Ne II line strengthening and color evolution as traces of the explosion mechanism
Synthetic observables from tECSN models show slower early red-color decline due to higher Ti/Cr and a late-time 12.8 μm Ne II line that strengthens over time, unlike comparable CO deflagration models.
-
Probabilistic Spectral Reconstruction of Trans-Neptunian Objects from Sparse Photometry: A Framework for Taxonomy, Survey Optimization, and Outlier Detection
Probabilistic PCA latent-space model with Bayesian inference reconstructs TNO near-IR spectra from photometry, achieving 95% credible-interval coverage and supporting taxonomy plus survey optimization.
-
A Census of Na D-traced neutral ISM and outflows at $0.6<z<4$
A JWST census detects neutral ISM absorption in 76 of 309 galaxies at 0.6<z<4 and outflows in 26, indicating AGN-driven neutral outflows dominate in quiescent systems at cosmic noon.
-
AMIGO: a Data-Driven Calibration of the JWST Interferometer
AMIGO is an end-to-end differentiable forward model of JWST AMI that corrects detector systematics to recover high-precision astrometry and detect close high-contrast companions.
-
Black-hole mass estimation through accretion disk spectral fitting for high-redshift blazars
MCMC fits to accretion disk SEDs of 23 high-z blazars give black hole masses of 10^8-10^10 solar masses and Eddington ratios 0.04-1, showing that ignoring IGM attenuation overestimates masses with larger bias at highe...
-
emcee: The MCMC Hammer
emcee delivers a stable Python implementation of the affine-invariant ensemble MCMC algorithm that requires minimal hand-tuning and supports easy parallelization.
-
Evidence for protostellar jets as a population of hadronic gamma-ray sources
Statistically significant gamma-ray detection from young stellar objects indicates hadronic particle acceleration in protostellar jets with output scaling to bolometric luminosity.
-
Probabilistic Spectral Reconstruction of Trans-Neptunian Objects from Sparse Photometry: A Framework for Taxonomy, Survey Optimization, and Outlier Detection
A PCA-based latent space model with Bayesian reconstruction achieves 95% credible interval coverage for TNO spectra from photometry using 4-10 components.
-
Probabilistic Analysis of Event-Mode Experimental Data
Direct probabilistic modeling of raw event-mode scattering data claims greater efficiency and lower systematic error than histogram-plus-least-squares methods.
-
A 3D Dust Map Based on Gaia, Pan-STARRS 1 and 2MASS
A new 3D dust reddening map with finer distance resolution, a spatial correlation prior, and Gaia-based distances covering the sky north of -30 degrees declination out to several kiloparsecs.
-
Inferring the star-formation histories of massive quiescent galaxies with BAGPIPES: Evidence for multiple quenching mechanisms
BAGPIPES fitting of 9289 massive quiescent galaxies shows most SFHs rise gradually then quench in 1-2 Gyr, with faster quenching at z>1 and slower at z<1, interpreted as multiple AGN feedback and gas-supply mechanisms.
-
A Spitzer Space Telescope Exploration Science Program to Search for Y Dwarf Variability
Spitzer mid-IR monitoring of all known Y dwarfs yields variability fractions of 35-75% that, when combined with L/T dwarf data, weakly support condensate cloud variations as the driver of brown dwarf variability.
-
The Importance of Galaxy-Wide Star Formation in Driving Winds at z~1
Observational analysis of 86 z~1 galaxies shows winds correlate with galaxy-wide SFR and Σ_SFR, not compact regions, implying distributed star formation drives outflows.
-
JWST Advanced Deep Extragalactic Survey (JADES) Data Release 5: stellar population catalogue for galaxies in GOODS-N and GOODS-S
JADES DR5 delivers a public catalog of Bayesian-inferred stellar masses, SFRs, SFHs, dust, metallicities, and AGN contributions for ~500k galaxies via Prospector with an evolving SFMS prior.
-
The DESIRED electron temperature relations in star-forming regions of the local Universe
A homogeneous analysis of 699 spectra yields Te-Te relations for ions including N II, O II, O III, S II, S III and Ar III, with lower dispersions for relations involving Te([N II]).
-
The DESIRED electron temperature relations in star-forming regions of the local Universe
Large-sample analysis finds Te([N II]) relations have lower dispersion than those for [O II] or [S II], making it a more reliable low-ionization temperature proxy.
-
Constraining the Molecular Kennicutt-Schmidt Relation with Multi-Transition CO Observations of Nearby Galaxies
Multi-transition CO observations reveal that the star formation-molecular gas relation becomes more linear for denser gas tracers, implying a volume density power-law index of approximately 1.5.
-
Detectability of deuterium in spectra of early-type stars
Synthetic spectra and MCMC analysis establish deuterium detection limits in early-type stars as a function of SNR, Teff, and v sin i, with upper limits D/H < -5.5 dex for HD 32115 and < -4.9 dex for 21 Peg.
-
Gravitational Wave Propagation in K-essence Cosmology: Theory and Observational Constraints
K-essence cosmology induces a redshift-dependent effective mass on gravitational waves, causing phase shifts that link scalar field dynamics to GW observables without changing wave speed or luminosity distance.
-
Markov chain Monte Carlo (MCMC) based Likelihood Extraction of Chiral-Odd Compton Form Factors from Deeply Virtual Exclusive Experiments
MCMC-based joint likelihood extraction constrains chiral-odd CFFs from DVMP cross-sections and asymmetries.
-
Correlation Between Hard X-Ray and Cosmic Neutrino Sources: From Obscured AGN to Blazars
Bayesian calibration of neutrino-hard X-ray luminosity relation on six AGN shows seven blazars are consistent, with joint permutation test indicating non-random association at 3.23 sigma.
-
Correlation Between Hard X-Ray and Cosmic Neutrino Sources: From Obscured AGN to Blazars
Seven blazars are posterior-predictively consistent with the L_hX-L_ν relation calibrated on six AGN, with redshift-partial and flux-permutation tests indicating residual correlation beyond distance bias.
-
Stellar Population Inference with Prospector
Prospector is a flexible code for Bayesian inference of stellar population parameters from multi-wavelength photometry and spectroscopy via forward modeling and posterior sampling.
-
Observational Constraints on $f(Q,T)$ Gravity in the Presence of DBI-Essence Scalar Field
Derives background solutions for linear f(Q,T)=αQ+βT plus DBI field and reports MCMC posteriors from Hubble, BAO, and SNIa data that are consistent with late-time constraints.
discussion (0)
Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.