A Bayesian Meta-Analysis of the Effect of Alcohol Use on HCV-Treatment Outcomes with a Comparison of Resampling Methods to Assess Uncertainty in Parameter Estimates.
Abstract not provided.
SEARCH · Search NASA
Search indexed NASA NTRS and DOE OSTI research on propulsion, heat transfer, battery materials and energy systems. Follow report and document links to the original sources.
Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.
Abstract not provided.
The uncertainty of soil moisture (SM) retrievals from satellite brightness temperature (TB) observations depends primarily on the choice of radiative transfer model (RTM) parameters, prior SM information and TB inputs. This paper studies the sensitivity of several (quasi-)operational and experimental SM retrieval products from the Soil Moisture Ocean Salinity (SMOS) mission to these choices at 11 reference sites, located in 7 watersheds across the United States (US). Different literature-based RTM parameter sets cause large biases between retrievals. Whereas typical RTM parameter sets are calibrated for SM retrievals, it is shown that a parameter set carefully optimized for TB forward modeling can also be used for retrieving SM. It is also shown that the inclusion of dynamic prior SM estimates in a Bayesian retrieval scheme can strongly improve SM retrievals, regardless of the choice of RTM parameters, and that the use of multi-angular and multi-polarization TB does not necessarily lead to superior retrievals compared to retrievals based on TB data at a single incidence angle and polarization. The second part of this paper evaluates ensemble uncertainty metrics for SM retrievals obtained by propagating a wide range of RTM parameters through the RTM. As expected for bounded variables, the spread in the ensemble SM retrievals is smallest for wet and dry SM values and highest for intermediate SM values. After removal of the strong long-term mean bias associated with the RTM parameter values for individual ensemble members, the remaining anomaly ensemble SM spread of 0.037 cu m/cu m approximates the actual time series unbiased root-mean-square-difference of 0.042 cu m/cu m between ensemble mean retrievals and in situ reference data across the reference sites. However, the temporal variability in the anomaly ensemble spread reveals higher-order biases in the retrieval error, which should be accounted for when characterizing retrieval error.
Data Science methods (many that are Bayesian based) are widely used in the physical sciences to estimate model parameters from experimental data, synthesize heterogeneous data, calibrate models, design experiments, and determine statistical significance of data. These methods provide a wealth of advantages over traditional analysis techniques because: 1) uncertainties are rigorously defined and propagated naturally through complex systems including covariance, 2) prior information is captured within the analysis framework (including rad-MHD and rad-hydro simulations), 3) competing models can be selected and/or ruled out using quantitative criteria, and 4) complex, heterogeneous data can be incorporated simultaneously. While these methods have been widely adopted as the gold standard in fields such as particle physics, astronomy, and biology, they have been slow to catch on in Inertial Confinement Fusion (ICF) and High Energy Density Physics (HEDP) research. Recently, several teams at LLNL, SNL, LANL, and the LLE have been exploring the use of these tools in their research and have found success. Here we propose that a concerted effort to consolidate these independent research efforts by developing and deploying common tools for use across the complex can revolutionize the way we approach data analysis, assimilation of theory and experiment, and decision making. The Bayesian formalism provides a means to accomplish this, but we are lacking certain infrastructure to make it happen on a large scale. Furthermore, once adopted, these techniques can be used to develop standards by which discoveries can be judged, similar to the so-called 5σ rule in high energy particle physics. Such standards may be used in the future to address the issue of unknown reproducibility in ICF and HED experiments caused by low shot rate and high cost per experiment. Our goals as a group are to advance the state of the art in HED measurement science by enabling: 1) better inferences from data with well-defined uncertainties, 2) better use of the data we have and continue to collect, 3) intelligent synthesis of data, 4) evaluation of the statistical significance of our data, and 5) informed decision making regarding the design of new experiments and instruments.
We present a newly developed version of BayeSED, a general Bayesian approach to the spectral energy distribution (SED) fitting of galaxies. The new BayeSED code has been systematically tested on a mock sample of galaxies. The comparison between the estimated and input values of the parameters shows that BayeSED can recover the physical parameters of galaxies reasonably well. We then applied BayeSED to interpret the SEDs of a large K{sub s} -selected sample of galaxies in the COSMOS/UltraVISTA field with stellar population synthesis models. Using the new BayeSED code, a Bayesian model comparison of stellar population synthesis models has been performed for the first time. We found that the 2003 model by Bruzual and Charlot, statistically speaking, has greater Bayesian evidence than the 2005 model by Maraston for the K{sub s} -selected sample. In addition, while setting the stellar metallicity as a free parameter obviously increases the Bayesian evidence of both models, varying the initial mass function has a notable effect only on the Maraston model. Meanwhile, the physical parameters estimated with BayeSED are found to be generally consistent with those obtained using the popular grid-based FAST code, while the former parameters exhibit more natural distributions. Based on the estimated physical parameters of the galaxies in the sample, we qualitatively classified the galaxies in the sample into five populations that may represent galaxies at different evolution stages or in different environments. We conclude that BayeSED could be a reliable and powerful tool for investigating the formation and evolution of galaxies from the rich multi-wavelength observations currently available. A binary version of the BayeSED code parallelized with Message Passing Interface is publicly available at https://bitbucket.org/hanyk/bayesed.
The Organization for Economic Cooperation and Development (OECD) Working Party on Nucelar Criticality Safety (WPNCS) has proposed a benchmark exercise representative of neutronic behavior in criticality experiments. Here, the goal is to develop confidence in data assimilation techniques used to adjust nuclear data. Participants are given synthetic experimental models with associated measured data and asked to estimate the model parameters given the model and measurements as well as provide predictions for separate application models. In this work, we performed data assimilation using Bayesian inverse Uncertainty Quantification (UQ) with machine learning surrogate models to produce posterior parameter distributions for the requested parameters and posterior predictive distributions for the requested responses. Several experimental models are shown to insufficiently inform the posterior parameter distributions for the applications involved. However, given sufficient experimental data, posterior parameter estimates yielded reduced uncertainty in the response predictions of interest while covering the experimental data.
Global Sensitivity Analysis (GSA) aims to understand the relative importance of uncertain input variables to model response. Conventional GSA involves calculating sensitivity (Sobol’) indices for a model with known model parameter distributions. However, model parameters are affected by aleatory and epistemic uncertainty, with the latter often caused by lack of data. In this paper, we propose a new framework to quantify uncertainty in probability model-form and model parameters resulting from small datasets and integrate these uncertainties into Sobol’ index estimates. First, the process establishes, through Bayesian multimodel inference, a set of candidate probability models and their associated probabilities. Imprecise Sobol’ indices are calculated from these probability models using an importance sampling reweighting approach. This results in probabilistic Sobol’ indices, whose distribution characterizes uncertainty in the sensitivity resulting from small dataset size. The imprecise Sobol’ indices thus provide a measure of confidence in the sensitivity estimate and, moreover, can be used to inform data collection efforts targeted to minimize the impact of uncertainties. Through an example studying the parameters of a Timoshenko beam, we show that these probabilistic Sobol’ indices converge to the true/deterministic Sobol’ indices as the dataset size increases and hence, distribution-form uncertainty reduces. The approach is then applied to assess the sensitivity of the out-of-plane properties of an E-glass fiber composite material to its constituent properties. This second example illustrates the approach for an important class of materials with wide-ranging applications when data may be lacking for some input parameters.
We report on the analysis of the high microlensing event MOA-2008-BLG-379, which has a strong microlensing anomaly at its peak due to a massive planet with a mass ratio of q = 6.9 × 10{sup –3}. Because the faint source star crosses the large resonant caustic, the planetary signal dominates the light curve. This is unusual for planetary microlensing events, and as a result, the planetary nature of this light curve was not immediately noticed. The planetary nature of the event was found when the Microlensing Observations in Astrophysics (MOA) Collaboration conducted a systematic study of binary microlensing events previously identified by the MOA alert system. We have conducted a Bayesian analysis based on a standard Galactic model to estimate the physical parameters of the lens system. This yields a host star mass of M{sub L}=3.3{sub −1.2}{sup +1.7} M{sub ⊙} orbited by a planet of mass m{sub P}=0.56{sub −0.27}{sup +0.24} M{sub Jup} at an orbital separation of a=3.3{sub −1.2}{sup +1.3} AU at a distance of D{sub L}=4.1{sub −1.9}{sup +1.7} kpc. The faint source magnitude of I {sub S} = 21.30 and relatively high lens-source relative proper motion of μ{sub rel} = 7.6 ± 1.6 mas yr{sup –1} imply that high angular resolution adaptive optics or Hubble Space Telescope observations are likely to be able to detect the source star, which would determine the masses and distance of the planet and its host star.
The Savannah River Site (SRS) spring operated pressure relief valve (SORV) maintenance intervals were evaluated using an approach provided by the American Petroleum Institute (API RP 581) for risk-based inspection technology (RBI). In addition, the impact of extending the inspection schedule was evaluated using Monte Carlo Simulation (MCS). The API RP 581 approach is characterized as a Weibull analysis with modified Bayesian updating provided by SRS SORV proof testing experience. Initial Weibull parameter estimates were updated as per SRS's historical proof test records contained in the Center for Chemical Process Safety (CCPS) Process Equipment Reliability Database (PERD). The API RP 581 methodology was used to estimate the SORV's probability of failing on demand (PFD), and the annual expected risk. The API RP 581 methodology indicates that the current SRS maintenance plan is conservative. Cost savings may be attained in certain mild service applications that present low PFD and overall risk. Current practices are reviewed and recommendations are made for extending inspection intervals. The paper gives an illustration of the inspection costs versus the associated risks by using API RP 581 Risk Based Inspection (RBI) Technology. A cost effective maintenance frequency balancing both financial risk and inspection cost is demonstrated.
We present BayeSED, a general purpose tool for Bayesian analysis of spectral energy distributions (SEDs) using pre-existing model SED libraries or their linear combinations. The artificial neural networks, principal component analysis, and multimodal-nested sampling (MultiNest) techniques are employed to allow the highly efficient sampling of posterior distribution and the calculation of Bayesian evidence. As a demonstration, we apply this tool to a sample of hyperluminous infrared galaxies (HLIRGs). The Bayesian evidence obtained for a pure starburst, a pure active galactic nucleus (AGN), and a linear combination of starburst+AGN models show that the starburst+AGN model has the highest evidence for all galaxies in this sample. The Bayesian evidence for the three models and the estimated contributions of starbursts and AGNs to infrared luminosity show that HLIRGs can be classified into two groups: one dominated by starbursts and the other dominated by AGNs. Other parameters and corresponding uncertainties about starbursts and AGNs are also estimated using the model with the highest Bayesian evidence. We find that the starburst region of the HLIRGs dominated by starbursts tends to be more compact and has a higher fraction of OB stars than that of HLIRGs dominated by AGNs. Meanwhile, the AGN torus of the HLIRGs dominated by AGNs tends to be more dusty than that of HLIRGs dominated by starbursts. These results are consistent with previous researches, but need to be tested further with larger samples. Overall, we believe that BayeSED could be a reliable and efficient tool for exploring the nature of complex systems such as dust-obscured starburst-AGN composite systems by decoding their SEDs.
We consider the problem of estimating the uncertainty in large-scale linear statistical inverse problems with high-dimensional parameter spaces within the framework of Bayesian inference. When the noise and prior probability densities are Gaussian, the solution to the inverse problem is also Gaussian, and is thus characterized by the mean and covariance matrix of the posterior probability density. Unfortunately, explicitly computing the posterior covariance matrix requires as many forward solutions as there are parameters, and is thus prohibitive when the forward problem is expensive and the parameter dimension is large. However, for many ill-posed inverse problems, the Hessian matrix of the data misfit term has a spectrum that collapses rapidly to zero. We present a fast method for computation of an approximation to the posterior covariance that exploits the lowrank structure of the preconditioned (by the prior covariance) Hessian of the data misfit. Analysis of an infinite-dimensional model convection-diffusion problem, and numerical experiments on large-scale 3D convection-diffusion inverse problems with up to 1.5 million parameters, demonstrate that the number of forward PDE solves required for an accurate low-rank approximation is independent of the problem dimension. This permits scalable estimation of the uncertainty in large-scale ill-posed linear inverse problems at a small multiple (independent of the problem dimension) of the cost of solving the forward problem.
This paper analyzes hierarchical Bayesian inverse problems using techniques from highdimensional statistics. Furthermore, our analysis leverages a property of hierarchical Bayesian regularizers that we call approximate decomposability to obtain non-asymptotic bounds on the reconstruction error attained by maximum a posteriori estimators. The new theory explains how hierarchical Bayesian models that exploit sparsity, group sparsity, and sparse representations of the unknown parameter can achieve accurate reconstructions in high-dimensional settings.
In this article, we propose a general Bayesian inference approach to the step-stress accelerated life test with type II censoring. We assume that the failure times at each stress level are exponentially distributed and the test units are tested in an increasing order of stress levels. We formulate the prior distribution of the parameters of life-stress function and integrate the engineering knowledge of product failure rate and acceleration factor into the prior. The posterior distribution and the point estimates for the parameters of interest are provided. Through the Markov chain Monte Carlo technique, we demonstrate a nonconjugate prior case using an industrial example. It is shown that with the Bayesian approach, the statistical precision of parameter estimation is improved and, consequently, the required number of failures could be reduced.
We wish to better constrain the properties of solar flares by exploring how parameterized models of solar flares interact with uncertainty estimation methods. We compare four different methods of calculating uncertainty estimates in fitting parameterized models to Ramaty High Energy Solar Spectroscopic Imager X-ray spectra, considering only statistical sources of error. Three of the four methods are based on estimating the scale-size of the minimum in a hypersurface formed by the weighted sum of the squares of the differences between the model fit and the data as a function of the fit parameters, and are implemented as commonly practiced. The fourth method is also based on the difference between the data and the model, but instead uses Bayesian data analysis and Markov chain Monte Carlo (MCMC) techniques to calculate an uncertainty estimate. Two flare spectra are modeled: one from the Geostationary Operational Environmental Satellite X1.3 class flare of 2005 January 19, and the other from the X4.8 flare of 2002 July 23.We find that the four methods give approximately the same uncertainty estimates for the 2005 January 19 spectral fit parameters, but lead to very different uncertainty estimates for the 2002 July 23 spectral fit. This is because each method implements different analyses of the hypersurface, yielding method-dependent results that can differ greatly depending on the shape of the hypersurface. The hypersurface arising from the 2005 January 19 analysis is consistent with a normal distribution; therefore, the assumptions behind the three non- Bayesian uncertainty estimation methods are satisfied and similar estimates are found. The 2002 July 23 analysis shows that the hypersurface is not consistent with a normal distribution, indicating that the assumptions behind the three non-Bayesian uncertainty estimation methods are not satisfied, leading to differing estimates of the uncertainty. We find that the shape of the hypersurface is crucial in understanding the output from each uncertainty estimation technique, and that a crucial factor determining the shape of hypersurface is the location of the low-energy cutoff relative to energies where the thermal emission dominates. The Bayesian/MCMC approach also allows us to provide detailed information on probable values of the low-energy cutoff, Ec, a crucial parameter in defining the energy content of the flare-accelerated electrons. We show that for the 2002 July 23 flare data, there is a 95% probability that Ec lies below approximately 40 keV, and a 68% probability that it lies in the range 7-36 keV. Further, the low-energy cutoff is more likely to be in the range 25-35 keV than in any other 10 keV wide energy range. The low-energy cutoff for the 2005 January 19 flare is more tightly constrained to 107 +/- 4 keV with 68% probability.
In this paper we consider the issue of paradigm evaluation by applying Bayes' theorem along the following nested hierarchy of progressively more complex structures: i) parameter estimation (within a model), ii) model selection and comparison (within a paradigm), iii) paradigm evaluation. In such a hierarchy the Bayesian evidence works both as the posterior's normalization at a given level and as the likelihood function at the next level up. Whilst raising no objections to the standard application of the procedure at the two lowest levels, we argue that it should receive a considerable modification when evaluating paradigms, when testability and fitting data are equally important. By considering toy models we illustrate how models and paradigms that are difficult to falsify are always favoured by the Bayes factor. We argue that the evidence for a paradigm should not only be high for a given dataset, but exceptional with respect to what it would have been, had the data been different. With this motivation we propose a measure which we term predictivity , as well as a prior to be incorporated into the Bayesian framework, penalising unpredictivity as much as not fitting data. We apply this measure to inflation seen as a whole, and to a scenario where a specific inflationary model is hypothetically deemed as the only one viable as a result of information alien to cosmology (e.g. Solar System gravity experiments, or particle physics input). We conclude that cosmic inflation is currently hard to falsify, but that this could change were external/additional information to cosmology to select one of its many models. We also compare this state of affairs to bimetric varying speed of light cosmology.
Understanding the physical conditions of dark molecular clouds and star-forming regions is an inverse problem subject to complicated chemistry that varies nonlinearly with both time and the physical environment. In this paper, we apply a Bayesian approach based on a Markov chain Monte Carlo (MCMC) method for solving the nonlinear inverse problems encountered in astrochemical modeling. We use observations for ice and gas species in dark molecular clouds and a time-dependent, gas-grain chemical model to infer the values of the physical and chemical parameters that characterize quiescent regions of molecular clouds. We show evidence that in high-dimensional problems, MCMC algorithms provide a more efficient and complete solution than more classical strategies. The results of our MCMC method enable us to derive statistical estimates and uncertainties for the physical parameters of interest as a result of the Bayesian treatment.
We wish to better constrain the properties of solar flares by exploring how parameterized models of solar flares interact with uncertainty estimation methods. We compare four different methods of calculating uncertainty estimates in fitting parameterized models to Ramaty High Energy Solar Spectroscopic Imager X-ray spectra, considering only statistical sources of error. Three of the four methods are based on estimating the scale-size of the minimum in a hypersurface formed by the weighted sum of the squares of the differences between the model fit and the data as a function of the fit parameters, and are implemented as commonly practiced. The fourth method is also based on the difference between the data and the model, but instead uses Bayesian data analysis and Markov chain Monte Carlo (MCMC) techniques to calculate an uncertainty estimate. Two flare spectra are modeled: one from the Geostationary Operational Environmental Satellite X1.3 class flare of 2005 January 19, and the other from the X4.8 flare of 2002 July 23. We find that the four methods give approximately the same uncertainty estimates for the 2005 January 19 spectral fit parameters, but lead to very different uncertainty estimates for the 2002 July 23 spectral fit. This is because each method implements different analyses of the hypersurface, yielding method-dependent results that can differ greatly depending on the shape of the hypersurface. The hypersurface arising from the 2005 January 19 analysis is consistent with a normal distribution; therefore, the assumptions behind the three non-Bayesian uncertainty estimation methods are satisfied and similar estimates are found. The 2002 July 23 analysis shows that the hypersurface is not consistent with a normal distribution, indicating that the assumptions behind the three non-Bayesian uncertainty estimation methods are not satisfied, leading to differing estimates of the uncertainty. We find that the shape of the hypersurface is crucial in understanding the output from each uncertainty estimation technique, and that a crucial factor determining the shape of hypersurface is the location of the low-energy cutoff relative to energies where the thermal emission dominates. The Bayesian/MCMC approach also allows us to provide detailed information on probable values of the low-energy cutoff, E{sub c} , a crucial parameter in defining the energy content of the flare-accelerated electrons. We show that for the 2002 July 23 flare data, there is a 95% probability that E{sub c} lies below approximately 40 keV, and a 68% probability that it lies in the range 7-36 keV. Further, the low-energy cutoff is more likely to be in the range 25-35 keV than in any other 10 keV wide energy range. The low-energy cutoff for the 2005 January 19 flare is more tightly constrained to 107 {+-} 4 keV with 68% probability. Using the Bayesian/MCMC approach, we also estimate for the first time probability density functions for the total number of flare-accelerated electrons and the energy they carry for each flare studied. For the 2002 July 23 event, these probability density functions are asymmetric with long tails orders of magnitude higher than the most probable value, caused by the poorly constrained value of the low-energy cutoff. The most probable electron power is estimated at 10{sup 28.1} erg s{sup -1}, with a 68% credible interval estimated at 10{sup 28.1}-10{sup 29.0} erg s{sup -1}, and a 95% credible interval estimated at 10{sup 28.0}-10{sup 30.2} erg s{sup -1}. For the 2005 January 19 flare spectrum, the probability density functions for the total number of flare-accelerated electrons and their energy are much more symmetric and narrow: the most probable electron power is estimated at 10{sup 27.66{+-}0.01} erg s{sup -1} (68% credible intervals). However, in this case the uncertainty due to systematic sources of error is estimated to dominate the uncertainty due to statistical sources of error.
The paper applies iterated and non-iterated extended Kalman filters to solve two practical meteorological inversion problems. Passive microwave satellite soundings are used to infer vertical temperature profiles and cloud parameters. In both cases it is shown that improvements can be obtained over presently used techniques. Finally, the results suggest that modern multivariate nonlinear recursive estimation techniques based in a Bayesian methodology can be a valuable tool in the area of remote sounding of atmospheric parameters.
Combining spatially-explicit long-term forest inventory and remotely sensed information from Light Detection and Ranging (LiDAR) datasets through statistical models can be a powerful tool for predicting and mapping above-ground biomass (AGB) at a range of geographic scales. We present and examine a novel modeling approach to improve prediction of AGB and estimate AGB growth using LiDAR data. The proposed model accommodates temporal misalignment between field measurements and remotely sensed data-a problem pervasive in such settings-by including multiple time-indexed measurements at plot locations to estimate AGB growth. We pursue a Bayesian modeling framework that allows for appropriately complex parameter associations and uncertainty propagation through to prediction. Specifically, we identify a space-varying coefficients model to predict and map AGB and its associated growth simultaneously. The proposed model is assessed using LiDAR data acquired from NASA Goddard's LiDAR, Hyper-spectral & Thermal imager and field inventory data from the Penobscot Experimental Forest in Bradley, Maine. The proposed model outperformed the time-invariant counterpart models in predictive performance as indicated by a substantial reduction in root mean squared error. The proposed model adequately accounts for temporal misalignment through the estimation of forest AGB growth and accommodates residual spatial dependence. Results from this analysis suggest that future AGB models informed using remotely sensed data, such as LiDAR, may be improved by adapting traditional modeling frameworks to account for temporal misalignment and spatial dependence using random effects.