Standard Bayesian analyses can be difficult to perform when the full likelihood, and consequently the full posterior distribution, is too complex or even impossible to specify or if robustness with respect to data or to model misspecifications is required. In these situations, we suggest to resort to a posterior distribution for the parameter of interest based on proper scoring rules. Scoring rules are loss functions designed to measure the quality of a probability distribution for a random variable, given its observed value. Important examples are the Tsallis score and the Hyvärinen score, which allow us to deal with model misspecifications or with complex models. Also the full and the composite likelihoods are both special instances of scoring rules.The aim of this paper is twofold. Firstly, we discuss the use of scoring rules in the Bayes formula in order to compute a posterior distribution, named SR-posterior distribution, and we derive its asymptotic normality. Secondly, we propose a procedure for building default priors for the unknown parameter of interest, that can be used to update the information provided by the scoring rule in the SR-posterior distribution. In particular, a reference prior is obtained by maximizing the average α−divergence from the SR-posterior distribution. For 0 ≤ |α| < 1, the result is a Jeffreys-type prior that is proportional to the square root of the determinant of the Godambe information matrix associated to the scoring rule. Some examples are discussed.
Both Approximate Bayesian Computation (ABC) and composite likelihood methods are useful for Bayesian and frequentist inference, respectively, when the likelihood function is intractable. We propose to use composite likelihood score functions as summary statistics in ABC in order to obtain accurate approximations to the posterior distribution. This is motivated by the use of the score function of the full likelihood, and extended to general unbiased estimating functions in complex models. Moreover, we show that if the composite score is suitably standardised, the resulting ABC procedure is invariant to reparameterisations and automatically adjusts the curvature of the composite likelihood, and of the corresponding posterior distribution. The method is illustrated through examples with simulated data, and an application to modelling of spatial extreme rainfall data is discussed.
Statistical applications often involve the calculation of intractable multidimensional integrals. The Laplace formula is widely used to approximate such integrals. However, in high-dimensional or small sample size problems, the shape of the integrand function may be far from that of the Gaussian density, and thus the standard Laplace approximation can be inaccurate. We propose an improved Laplace approximation that reduces the asymptotic error of the standard Laplace formula by one order of magnitude, thus leading to third-order accuracy. We also show, by means of practical examples of various complexity, that the proposed method is extremely accurate, even in high dimensions, improving over the standard Laplace formula. Such examples also demonstrate that the accuracy of the proposed method is comparable with that of other existing methods, which are computationally more demanding. An R implementation of the improved Laplace approximation is also provided through the R package iLaplace available on CRAN.
We explore the use of higher-order tail area approximations for Bayesian simulation. These approximations give rise to an alternative simulation scheme to MCMC for Bayesian computation of marginal posterior distributions for a scalar parameter of interest, in the presence of nuisance parameters. Its advantage over MCMC methods is that samples are drawn independently with lower computational time and the implementation requires only standard maximum likelihood routines. The method is illustrated by a genetic linkage model, a normal regression with censored data and a logistic regression model. approximate (1); see, e.g., Robert and Casella (2004). All these techniques use simulation to avoid tailoring analytical work to specific models. However, these methods may present some difficulties, especially when d is large, and may have poor tail behaviour.Parallel with these developments has been the development of analytical higher-order approximations for parametric inference in small sample (see, e.g., Brazzale and Davison, 2008, and references therein). Using higher-order asymptotics it is possible to avoid the difficulties related to numerical methods and obtain accurate approximations of (1) and related quantities (see, e.g., Reid, 1996Reid, , 2003Sweeting, 1996, andBrazzale et al., 2007). These methods are highly accurate in many situations, but are nevertheless underused compared to simulation-based procedures (Brazzale and Davison, 2008).Starting from higher-order tail area approximations for (1) (see Diciccio and Martin, 1991;Reid, 1996and Brazzale et al., 2007, this paper describes the implementation and the use of a sampling scheme that give rise to very accurate computation of marginal posterior densities, and related quantities, such as posterior summaries. The implementation of the proposed higher-order tail area approximation (HOTA) sampling scheme is available at little additional computation cost over simple first-order approximations, and it has the advantage over MCMC methods that samples are drawn independently in much lower computation time. This technique applies to regular models only; precise regularity conditions for their validity are given in Kass et al. (1990).The paper is organized as follows. Section 2 briefly reviews higher-order approximations for the marginal posterior distribution (1), and for the corresponding tail area. Section 3 describes the proposed HOTA sampling scheme and its implementation. Numerical examples and applications are discussed in Section 4. Finally, some concluding remarks are given in Section 5.
Summary We discuss an approach of robust fitting on nonlinear regression models, both in a frequentist and a Bayesian approach, which can be employed to model and predict the contagion dynamics of COVID‐19 in Italy. The focus is on the analysis of epidemic data using robust dose‐response curves, but the functionality is applicable to arbitrary nonlinear regression models.
Liquid biopsy represents a new frontier of cancer diagnosis and prognosis, which allows the isolation of tumor cells released in the blood stream. The extremely low abundance of these cells needs appropriate methodologies for their identification and enumeration. Herein we present a new protocol based on surface enhanced resonance Raman scattering (SERRS) gold multivalent nanostructures to identify and enumerate tumor cells with epithelial and mesenchimal markers. The validation of the protocol is obtained with spiked samples of peripheral blood mononuclear cells (PBMC). Gold nanostructures are functionalized with SERRS labels and with antibodies to link the tumor cells. Three types of such nanosystems were simultaneously used and the protocol allows obtaining the identification of all individual tumor cells with the help of a Random Forest ensemble learning method.
Glucose-regulated protein94 (Grp94), the most represented endoplasmic reticulum (ER)-resident heat shock protein (HSP), is a tumor antigen shared by different types of solid and hematological tumors. The tumor-specific feature of Grp94 is its translocation from the ER to the cell surface where it displays pro-oncogenic functions. This un-physiological location has important implications for both the tumor pathology and anti-tumor therapy. We wanted to address the question of whether Grp94 could be measured as liquid marker in cancer patients in order to make predictions of diagnostic and therapeutic relevance for the tumor. To this aim, we performed an in-depth investigation on patients with primary tumors of the gastrointestinal (GI) tract, using different methodological approaches to detect Grp94 in tumor tissues, plasma and peripheral blood mononuclear cells (PBMCs). Results indicate that Grp94 is not only the antigen highly expressed in any tumor tissue and in cells of tumor infiltrates, mostly B lymphocytes, but it is also found in the circulation. However, the only form in which Grp94 was detected in the plasma of any patients and in B lymphocytes induced to proliferate, was that of stable complexes with Immunoglobulin (Ig)G. Using a specific immune-enzyme assay to measure plasma Grp94-IgG complexes, we showed that Grp94-IgG complexes were significantly increased in cancer patients compared to healthy control subjects, serving as diagnostic tumor biomarker. Results also demonstrate that the stimulation of patient PBMCs with Grp94-IgG complexes led to an increased secretion of inflammatory cytokines that might drive a potentially beneficial anti-tumor effect.
We discuss an approach for deriving robust posterior distributions from M -estimating functions using Approximate Bayesian Computation (ABC) methods. In particular, we use M -estimating functions to construct suitable summary statistics in ABC algorithms.The theoretical properties of the robust posterior distributions are discussed. Special attention is given to the application of the method to linear mixed models. Simulation results and an application to a clinical study demonstrate the usefulness of the method.An R implementation is also provided in the robustBLME package.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.