Review
Engineering, Civil
Aditya Pandey, Ashmeet Singh, Paolo Gardoni
Summary: This paper reviews the diagrammatic perturbation theory, a technique in Information Field Theory, for analytically estimating moments of perturbative non-Gaussian distributions. When dealing with physical phenomena, which often exhibit non-Gaussian features, approximation of the underlying distribution and inference of its parameter form are commonly used. More rigorous analysis methods such as Markov Chain Monte Carlo can also be employed, but are computationally expensive.
Article
Business
Mike G. Tsionas
Summary: This paper presents results from Bayesian analysis of random switching exponential smoothing models, showing that the new methods are robust and easy to implement. Monte Carlo simulations and real data sets demonstrate the methods' strong performance, especially when extended with a Markov chain assumption on the slope of the trend. Model comparison and selection tools are also provided for out-of-sample behavior analysis.
TECHNOLOGICAL FORECASTING AND SOCIAL CHANGE
(2022)
Article
Statistics & Probability
Alice Martin, Marie-Pierre Etienne, Pierre Gloaguen, Sylvain Le Corff, Jimmy Olsson
Summary: This article proposes a new Sequential Monte Carlo algorithm to perform online estimation when certain densities are intractable in state space models. It introduces a pseudo-marginal backward importance sampling step to estimate expectations. The proposed algorithm significantly reduces computational time and broadens the class of eligible models. Its performance is assessed in different scenarios. Appendices for this article are available online.
JOURNAL OF COMPUTATIONAL AND GRAPHICAL STATISTICS
(2023)
Article
Computer Science, Interdisciplinary Applications
Alexandre L. M. Levada
Summary: This paper proposes an information-geometric framework to analyze Gaussian-Markov random fields (GMRF's) and measures the variations in the metric tensor components to understand the changes in the system over time. A method based on infinitesimal displacements using Markov Chain Monte Carlo simulations is used to compute distances between two systems operating in different regimes using the Fisher metric. Additionally, an expression for the KL-divergence between two GMRF models is derived, showing that it can be a good replacement for the Fisher information based distance. Moreover, the paper reveals an asymmetric pattern of evolution when the system moves towards different entropic states, indicating the emergence of an intrinsic notion of time based on the geometric properties of its parametric space.
JOURNAL OF COMPUTATIONAL SCIENCE
(2022)
Article
Computer Science, Artificial Intelligence
K. Vinodha, E. S. Gopi
Summary: This paper proposes a method that uses Ant Colony Optimization and Adaptive boosting to optimize hierarchical binary classifier structures, in order to improve the overall detection rate. The effectiveness of this method was demonstrated through Monte Carlo simulations and compared with the default structure, showing a significant improvement in detection rate. This method is suitable for applications that require hierarchical multi-level classification.
EXPERT SYSTEMS WITH APPLICATIONS
(2024)
Article
Environmental Sciences
Gerrit Schoups, Mohsen Nasseri
Summary: This study introduces a Bayesian hierarchical model that combines monthly water balance data and estimates data errors and error-corrected water balance components, leading to significant improvements in the accuracy of posterior water balance estimates.
WATER RESOURCES RESEARCH
(2021)
Article
Engineering, Electrical & Electronic
Jean-Baptiste Courbot, Bruno Colicchio
Summary: This paper discusses the problem of Bayesian deconvolution and proposes a new model TGRF, which obtains the latent field of GMRF through nonlinear transformation. We also propose a Bayesian inference method and utilize the expectation-maximization algorithm to jointly deconvolve and estimate the statistical model parameters of GMRF and TGRF. Numerical results validate the effectiveness of this method on different types of images.
IEEE SIGNAL PROCESSING LETTERS
(2022)
Article
Statistics & Probability
Santeri Karppinen, Sumeetpal S. S. Singh, Matti Vihola
Summary: Conditional particle filters (CPFs) with backward/ancestor sampling are powerful sampling methods for latent states of dynamic models. However, their performance deteriorates with weakly informative observations and slowly mixing dynamics. This article introduces resampling strategies and a generalization of CPF to address these complications. Practical tuning strategies for choosing appropriate blocking are also presented.
JOURNAL OF COMPUTATIONAL AND GRAPHICAL STATISTICS
(2023)
Article
Biochemical Research Methods
Zhuoran Ding, Marylyn D. Ritchie, Benjamin F. Voight, Wei-Ting Hwang
Summary: This study identifies causal factors for complex traits in humans using observational studies and Mendelian randomization experiments, and infers the effect of hidden mediators on the outcome trait through a mediation model framework. Simulation and data analysis show the effectiveness of the proposed method.
BMC BIOINFORMATICS
(2022)
Article
Computer Science, Interdisciplinary Applications
Mohammad Aminpour, Reza Alaie, Sajjad Khosravi, Navid Kardani, Sara Moridpour, Majidreza Nazem
Summary: This paper investigates the efficiency of machine learning models and Artificial Neural Networks in predicting the results of large datasets for random field slope stability analysis. The study finds that the models perform well in predicting the probability of failure, and their efficiency is further enhanced by using the proposed probability summation method. The models also prove to be efficient in predicting the factor of safety for stratified random field anisotropic heterogeneous slopes.
COMPUTERS AND GEOTECHNICS
(2023)
Article
Materials Science, Multidisciplinary
Mina Karimi, Mehrdad Massoudi, Kaushik Dayal, Matteo Pozzi
Summary: We explore the use of Bayesian inference for solving large-scale inverse problems governed by partial differential equations (PDEs). Markov Chain Monte Carlo (MCMC) methods are employed to generate samples and estimate posterior uncertainty. To improve efficiency in high-dimensional problems, the gradient and Hessian of the target potential are utilized through Hamiltonian Monte Carlo (HMC). Specifically, we apply this framework to infer the field of soil permeability from pore pressure observations using a nonlinear PDE poromechanics model. The performance of different sampling approaches, as well as the impact of dimensionality and non-Gaussian distributions, are studied.
MATHEMATICS AND MECHANICS OF SOLIDS
(2023)
Article
Engineering, Mechanical
P. L. Green, L. J. Devlin, R. E. Moore, R. J. Jackson, J. Li, S. Maskell
Summary: This paper discusses the optimization of the 'L-kernel' in Sequential Monte Carlo samplers to improve performance, resulting in reduced variance of estimates and fewer resampling requirements.
MECHANICAL SYSTEMS AND SIGNAL PROCESSING
(2022)
Article
Computer Science, Interdisciplinary Applications
DanHua ShangGuan
Summary: The Monte Carlo method is a powerful tool in many research fields, but the increasing complexity of physical models and mathematical models requires efficient algorithms to overcome the computational cost.
JOURNAL OF COMPUTATIONAL PHYSICS
(2021)
Article
Engineering, Mechanical
Adolphus Lye, Alice Cicirello, Edoardo Patelli
Summary: This tutorial paper reviews the use of advanced Monte Carlo sampling methods in Bayesian model updating for engineering applications, introducing different methods and comparing their performance. Three case studies demonstrate the advantages and limitations of these sampling techniques in parameter identification, posterior distribution sampling, and stochastic identification of model parameters.
MECHANICAL SYSTEMS AND SIGNAL PROCESSING
(2021)
Article
Geosciences, Multidisciplinary
Martin Outzen Berild, Geir-Arne Fuglstad
Summary: Isotropic covariance structures may not be appropriate for phenomena in three-dimensional spaces. We propose a non-stationary anisotropic Gaussian random fields (GRFs) model in three dimensions using stochastic partial differential equations (SPDEs), with efficient computations based on Gaussian Markov random field approximations. The unique aspect of our model lies in the parameterization of spatially varying anisotropy through vector fields.
SPATIAL STATISTICS
(2023)