Article
Biology
Yichao Li, Wenshuo Wang, K. E. Deng, Jun S. Liu
Summary: Sequential Monte Carlo algorithms are powerful tools for inference with dynamical systems. Resampling is a key step in the algorithm, with different strategies such as stratified resampling and optimal transport resampling used in practice. This study shows the equivalence of optimal transport resampling and stratified resampling in one-dimensional cases and demonstrates an improved variance rate for stratified resampling in general d-dimensional cases. The study also presents bounds on the Wasserstein distance and convergence rates for sequential quasi-Monte Carlo with Hilbert curve resampling.
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
Computer Science, Interdisciplinary Applications
Aaron Myers, Alexandre H. Thiery, Kainan Wang, Tan Bui-Thanh
Summary: The SET method generates approximate samples from a Bayesian posterior distribution by solving a sequence of optimal transport problems. It converges weakly to the true posterior as sample size approaches infinity. Compared to standard SMC methods, SET is more robust and requires less computational efforts when exploring complex posterior distributions.
JOURNAL OF COMPUTATIONAL PHYSICS
(2021)
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
Statistics & Probability
Hai-Dang Dau, Nicolas Chopin
Summary: The paper proposes a new waste-free sequential Monte Carlo (SMC) algorithm that utilizes the outputs of all intermediate Markov chain Monte Carlo (MCMC) steps as particles. The consistency and asymptotic normality of its output are established, and insights on estimating the asymptotic variance of any particle estimate are developed. Empirical results show that waste-free SMC tends to outperform standard SMC samplers, particularly in scenarios where the mixing of the considered MCMC kernels decreases across iterations.
JOURNAL OF THE ROYAL STATISTICAL SOCIETY SERIES B-STATISTICAL METHODOLOGY
(2022)
Article
Computer Science, Theory & Methods
Jiangqi Wu, Linjie Wen, Peter L. Green, Jinglai Li, Simon Maskell
Summary: Many real-world problems require estimation of parameters of interest in a Bayesian framework from sequentially collected data. Conventional methods for sampling from posterior distributions do not efficiently address these problems as they do not consider the sequential structure of the data. Therefore, sequential methods like EnKF and SMCS are often used to update the posterior distribution and solve such problems.
STATISTICS AND COMPUTING
(2022)
Article
Computer Science, Theory & Methods
Xinzhu Liang, Shangda Yang, Simon L. L. Cotter, Kody J. H. Law
Summary: This paper addresses the problem of estimating expectations when the normalizing constant of the target distribution is unknown and the unnormalized target needs to be approximated at finite resolution. Building upon a recently introduced multi-index sequential Monte Carlo (SMC) ratio estimator, this work combines the complexity improvements of multi-index Monte Carlo (MIMC) with the efficiency of SMC for inference. The proposed method uses a randomization strategy to remove bias entirely, simplifying the estimation process, particularly in the context of MIMC.
STATISTICS AND COMPUTING
(2023)
Article
Automation & Control Systems
Adrien Corenflos, Nicolas Chopin, Simo Sarkka
Summary: This study proposes a new particle smoother, dSMC, that can process a large number of observations with lower time complexity on parallel processors. The variance of the smoothing estimates computed by dSMC can be reduced by designing good proposal distributions and using lazy resampling.
JOURNAL OF MACHINE LEARNING RESEARCH
(2022)
Article
Computer Science, Interdisciplinary Applications
Nicholas Michaud, Perry de Valpine, Daniel Turek, Christopher J. Paciorek, Dao Nguyen
Summary: nimble is an R package that provides flexibility in model specification and allows users to program model-generic algorithms. nimbleSMC is an extension package that contains algorithms for state-space model analysis using SMC techniques.
JOURNAL OF STATISTICAL SOFTWARE
(2021)
Article
Economics
Nianling Wang, Zhusheng Lou
Summary: The stochastic volatility (SV) model is widely used to study time-varying volatility. However, the linearity assumption for transition equation in basic SV model is restrictive. To allow for nonlinearity, we proposed a semiparametric SV model that specifies a nonparametric transition equation for log-volatility using natural cubic splines. The empirical applications to Bitcoin and convertible bond return data indicate that the transition equations of their log-volatility are highly nonlinear. Taking nonlinearity into account, the semi-parametric SV model can improve the likelihood of the basic SV model both in-sample and out-of-sample.
ECONOMIC MODELLING
(2023)
Article
Computer Science, Interdisciplinary Applications
Kathryn Turnbull, Christopher Nemeth, Matthew Nunes, Tyler McCormick
Summary: Dynamic network data describes interactions among a fixed population over time. The latent space framework and sequential Monte Carlo (SMC) methods are utilized to model and estimate the parameters of dynamic latent space network models. The proposed approach offers scalability, convenience of update, and flexibility under model variants, as demonstrated through simulation and analysis of a real-world dataset.
COMPUTATIONAL STATISTICS & DATA ANALYSIS
(2023)
Article
Computer Science, Interdisciplinary Applications
Alexander Kreuzer, Luciana Dalla Valle, Claudia Czado
Summary: This research proposes a novel flexible class of multivariate nonlinear non-Gaussian state space models based on copulas. The observation equation and the state equation are defined by copula families that are not necessarily equal. Inference is performed using the Hamiltonian Monte Carlo method within the Bayesian framework. Simulation studies demonstrate that the copula-based approach is highly flexible, capturing a wide range of dependence structures and enabling handling of missing data. The application to atmospheric pollutant measurement data illustrates the model's suitability for accurate modeling and prediction in the presence of missing values. Comparisons with a Gaussian linear state space model and Bayesian additive regression trees show the superior predictive accuracy of the proposed model.
COMPUTATIONAL STATISTICS & DATA ANALYSIS
(2023)
Article
Engineering, Environmental
Marco Bacci, Jonas Sukys, Peter Reichert, Simone Ulzega, Carlo Albert
Summary: Due to limited knowledge about complex environmental systems, predicting their behavior under different scenarios or decision alternatives is uncertain. Considering, quantifying, and communicating this uncertainty is important for societal decisions. Stochastic models are often necessary to adequately describe uncertainty, but calibrating these models presents methodological and numerical challenges. To address this, we compare three numerical approaches and find that their performance is comparable for analyzing a stochastic hydrological model with hydrological data, suggesting that generality and practical considerations can guide technique choice for specific applications.
STOCHASTIC ENVIRONMENTAL RESEARCH AND RISK ASSESSMENT
(2023)
Article
Statistics & Probability
Joe Marion, Joseph Mathews, Scott C. Schmidler
Summary: We present bounds for the finite-sample error of sequential Monte Carlo samplers on static spaces. Our approach explicitly relates the performance of the algorithm to properties of the chosen sequence of distributions and mixing properties of the associated Markov kernels. This allows us to give the first finite-sample comparison to other Monte Carlo schemes. We obtain bounds for the complexity of sequential Monte Carlo approximations for a variety of target distributions such as finite spaces, product measures, and log-concave distributions including Bayesian logistic regression. The bounds obtained are within a logarithmic factor of similar bounds obtainable for Markov chain Monte Carlo.
ANNALS OF STATISTICS
(2023)
Article
Computer Science, Theory & Methods
Ajay Jasra, Kody J. H. Law, Neil Walton, Shangda Yang
Summary: This paper discusses the problem of estimating expectations with respect to a target distribution with an unknown normalizing constant. A multi-index sequential Monte Carlo method is proposed to improve the efficiency of inference, and it is illustrated on various examples to verify its effectiveness.
FOUNDATIONS OF COMPUTATIONAL MATHEMATICS
(2023)
Article
Ecology
Florian Lecorvaisier, Dominique Pontier, Benoit Soubeyrand, David Fouchet
Summary: Research has found that the use of vaccines that do not entirely block pathogen transmission may lead to the evolution of more virulent strains. High vaccine coverage favors the emergence and prevalence of avirulent strains, and competition between strains is crucial for the eradication of toxigenic strains when these vaccines are used.
ECOLOGICAL MODELLING
(2024)
Article
Ecology
Monica E. Barros, Ana Arriagada, Hugo Arancibia, Sergio Neira
Summary: The stock biomass of carrot prawn in the south-central area of Chile has decreased in the past 12 years, mainly due to fishing mortality. Predation mortality has been less studied and quantified, so it is important to estimate and compare predation and fishing mortality to understand their effects on fishing stocks. A food web model was built to analyze the biomass changes and evaluate the relative contribution of different mortality factors. The results showed that predation mortality was the main component of total mortality for carrot prawns and yellow prawns.
ECOLOGICAL MODELLING
(2024)
Article
Ecology
Shubham Krishna, Victoria Peterson, Luisa Listmann, Jana Hinners
Summary: This study incorporated viral dynamics into an ecosystem model to investigate the effects of viruses on ecosystem dynamics under current and future climatic conditions. The results showed that the presence of viruses increased nutrient retention in the upper water column, leading to a reduction in phytoplankton biomass and transfer of biomass to higher trophic levels.
ECOLOGICAL MODELLING
(2024)
Article
Ecology
Zahra Dehghan Manshadi, Parastoo Parivar, Ahad Sotoudeh, Ali Morovati Sharifabadi
Summary: This study demonstrates the importance of strategies such as limiting built-up areas, preserving green spaces, and protecting water resources on the urban carrying capacity in arid and semi-arid regions. Implementing a combination of policies aimed at enhancing urban green spaces and regulating water demand is found to be the most effective in terms of health and urban carrying capacity.
ECOLOGICAL MODELLING
(2024)
Article
Ecology
Shay S. Keretz, Daelyn A. Woolnough, Todd J. Morris, Edward F. Roseman, David T. Zanatta
Summary: This study surveyed native freshwater mussels in the St. Clair-Detroit River system and found 14 live unionids representing 9 species. However, the model used to predict their presence in the main channels was not successful. The study also revealed characteristic differences between the St. Clair and Detroit Rivers.
ECOLOGICAL MODELLING
(2024)
Article
Ecology
Zhengrong Zhang, Xuemei Li, Xinyu Liu, Kaixin Zhao
Summary: This study examines land use change in the Chinese Tianshan mountainous region using system dynamics and patch-generating land use simulation models. The results show an expansion in forest and construction land, a decline in grassland area, and an increase in cultivated land area from 2005 to 2020. By 2040, unused land, grassland, and water are expected to decrease while other land types increase, with construction land showing the most significant increase. The study provides insights for future ecological and environmental management in the region.
ECOLOGICAL MODELLING
(2024)
Article
Ecology
Amira Khelifa, Nadjia El Saadi
Summary: This paper develops an agent-based model to study malaria disease transmission, taking into account the interactions between hosts, vectors, and aquatic habitats, as well as their geographical locations. The simulation results highlight the significant role of aquatic habitats in infection transmission and disease persistence, and demonstrate the effectiveness of eliminating these habitats in limiting disease transmission.
ECOLOGICAL MODELLING
(2024)
Article
Ecology
Guillaume Peron
Summary: The theory for movement-based coexistence between species often overlooks small-scale, station-keeping movements. However, at this scale, there are many instances where positive correlations exist between species traits that are expected to be negatively correlated based on current theory. Through simulations, the researcher presents a counter-example to demonstrate that functional tradeoffs are not a necessary condition for movement-based coexistence. This study highlights the significance of species-specific space use patterns under the time allocation tradeoff hypothesis.
ECOLOGICAL MODELLING
(2024)
Article
Ecology
Sandra Y. Mendiola, Nicole M. Gerardo, David J. Civitello
Summary: Research on the use of insect microbial symbionts as a means of controlling the spread of insect vectors and the pathogens they carry has made significant progress in the last decade. This study focused on the relative importance of simultaneous effects caused by a symbiont called Caballeronia spp. on the ability of squash bugs to transmit phytopathogenic Serratia marcescens. The researchers found that infection with Caballeronia significantly reduced pathogen titers and cleared S. marcescens in bugs, thus reducing the vectoring potential of these pests. The study also showed that maximizing symbiont prevalence in the vector population is crucial for effectively mitigating plant infections.
ECOLOGICAL MODELLING
(2024)
Article
Ecology
Shirui Hao, Dongryeol Ryu, Andrew W. Western, Eileen Perry, Heye Bogena, Harrie Jan Hendricks Franssen
Summary: This study investigates the sensitivity of model yield prediction to uncertainties in model parameters and inputs using the Sobol' method. The results show that yield is more sensitive to changes in water availability and nitrogen availability, depending on soil, management, and weather conditions.
ECOLOGICAL MODELLING
(2024)
Article
Ecology
Nitika Mundetia, Devesh Sharma, Aditya Sharma
Summary: This study focused on assessing groundwater sustainability using different modeling approaches in a river basin in Rajasthan, India. The results showed a decrease in future groundwater recharge and emphasized the need for better management and conservation practices to achieve sustainable development goals.
ECOLOGICAL MODELLING
(2024)
Article
Ecology
Sukdev Biswas, Sk Golam Mortoja, Ritesh Kumar Bera, Sabyasachi Bhattacharya
Summary: Bacteria play a crucial role in regulating the nutrient cycle of ecosystems, and maintaining a thriving bacterial population is essential for the sustainability of these environments. This study introduces the concept of cooperation as a group defense mechanism employed by bacteria and incorporates it into the functional response, offering a more comprehensive understanding of the complex tritrophic food chain dynamics. The results highlight the importance of a balance between strong group defense and moderate cooperation for bacteria sustainability and overall system stability.
ECOLOGICAL MODELLING
(2024)
Article
Ecology
D. Z. M. Le Gouvello, S. Heye, L. R. Harris, J. Temple-Boyer, P. Gaspar, M. G. Hart-Davis, C. Louro, R. Nel
Summary: This study modeled the dispersal pathways and compared potential dispersal corridors of different sea turtle species in the Western Indian Ocean. The results showed that ocean currents play a major role in driving dispersal, with species and years exhibiting differences in dispersal patterns. Active swimming had little influence on dispersal during the first year.
ECOLOGICAL MODELLING
(2024)
Review
Ecology
Yingying Duan, Haina Rong, Gexiang Zhang, Sergey Gorbachev, Dunwu Qi, Luis Valencia-Cabrera, Mario J. Perez-Jimenez
Summary: Computing models are an effective way to study population dynamics of endangered species like giant pandas. This paper proposes a unified framework and conducts a comprehensive survey of computing models for giant panda ecosystems. Multi-factor computing models are more suitable for studying giant panda ecosystems.
ECOLOGICAL MODELLING
(2024)
Article
Ecology
Samantha Lai, Theophilus Zhi En Teo, Arief Rullyanto, Jeffery Low, Karenne Tun, Peter A. Todd, Siti Maryam Yaakub
Summary: Understanding the exchange of genetic material among populations in the marine environment is crucial for conservation efforts. Agent-based models are increasingly used to predict dispersal pathways, including for seagrasses. This study highlights the importance of considering both sexual propagules and asexual vegetative fragments when evaluating seagrass connectivity.
ECOLOGICAL MODELLING
(2024)