Fast and accurate Bayesian model criticism and conflict diagnostics using R‐INLA

Ferkingstad, Egil; Held, Leonhard; Rue, Håvard

doi:10.1002/sta4.163

Cited by 13 publications

(9 citation statements)

References 34 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Further open questions in this area include how to improve power to detect conflict, and how to make such methods more accessible by improving the computational feasibility of systematic conflict assessment. As with any cross-validatory framework, multiple node-splitting can be computationally burdensome, so for hierarchical models, (Ferkingstad et al, 2017) have proposed an INLA approach to fast conflict diagnostics. A final area of open research related to understanding the influence of different, potentially conflicting, evidence sources on inference is the adaptation of value of information methods to evidence synthesis (Jackson et al, 2017).…”

Section: Discussionmentioning

confidence: 99%

Analysing Multiple Epidemic Data Sources

Angelis¹,

Presanis²

2018

Preprint

View full text Add to dashboard Cite

Evidence-based knowledge of infectious disease burden, including prevalence, incidence, severity and transmission, in different population strata and locations, and possibly in real time, is crucial to the planning and evaluation of public health policies. Direct observation of a disease process is rarely possible. However, latent characteristics of an epidemic and its evolution can often be inferred from the synthesis of indirect information from various routine data sources, as well as expert opinion. The simultaneous synthesis of multiple data sources, often conveniently carried out in a Bayesian framework, poses a number of statistical and computational challenges: the heterogeneity in type, relevance and granularity of the data, together with selection and informative observation biases, lead to complex probabilistic models that are difficult to build and fit, and challenging to criticize. Using motivating case studies of influenza, this chapter illustrates the cycle of model development and criticism in the context of Bayesian evidence synthesis, highlighting the challenges of complex model building, computationally efficient inference, and conflicting evidence.

show abstract

Section: Discussionmentioning

confidence: 99%

Analysing Multiple Epidemic Data Sources

Angelis¹,

Presanis²

2018

Preprint

View full text Add to dashboard Cite

show abstract

“…To consider the potential influence of tree height and fire history on our results we re‐ran the model outlined above using tree height instead of tree canopy type and fire history instead of NMDS axis 2. Various assessment criteria for our global model suggested a good model fit, including a Watanabe–Akaike Information Criterion (a Bayesian equivalent of the well‐known AIC) of −14,484, Deviance Information Criterion −14,100 and conditional predictive ordination −3763 (Ferkingstad, Held, & Rue, ). Using the Bayesian framework, we assessed the support for the importance of each fixed factor in the model by examining the 95% credible intervals around the posterior mean.…”

Section: Methodsmentioning

confidence: 99%

Savannah trees buffer herbaceous plant biomass against wild and domestic herbivores

Smith

Graae

Bukombe

et al. 2020

Applied Vegetation Science

View full text Add to dashboard Cite

Questions Given the growing abundance and dominance of domestic herbivores in savannah ecosystems, can trees maintain plant herbaceous standing biomass under increasing herbivore pressure? Are there differences in the capacity of leguminous and non‐leguminous trees in sustaining understory herbaceous biomass? And finally, to what extent does plant community composition underneath trees modulate the effects of herbivore assemblage and abundance on herbaceous biomass? Location Pasturelands and protected areas along the borders of the Serengeti National Park, Tanzania, East Africa. Method Monthly herbaceous biomass was monitored non‐destructively using a calibrated pasture disc. Measurements were taken within a network of paired exclosures and open plots, underneath large leguminous and non‐leguminous tree canopies and outside canopies. Herbivore community assemblage compositions and abundances were characterised using monthly dung counts, and herbaceous plant community composition was surveyed annually every wet season over two years. Results Overall, we found that trees promote herbaceous standing biomass, particularly in the presence of moderate herbivory rather than under herbivore exclusion. Greater herbivore abundance and livestock dominance reduced herbaceous plant biomass, but trees, particularly leguminous trees, limited these negative effects. This capacity for trees to limit the effect of herbivores was related to herbaceous plant species composition. Understory plant communities that were compositionally typical of protected areas sustained the highest plant biomass when found in pasturelands with high herbivore pressure. Conclusion Our findings give greater credence to the importance of preserving large trees in savannah landscapes increasingly dominated by high abundances of livestock. Moreover, our results highlight that park managers and pastoralists need to maintain the specialist herbaceous understory community beneath trees in order to benefit from facilitative tree–understory interactions.

show abstract

“…We implemented hierarchical Bayesian spatiotemporal models with the r‐inla package (Rue et al, 2009, 2017), using the default and recommended settings for priors (Held, Schrödle, & Rue, 2010) and the ‘ laplace ’ (most accurate) approximation to estimate the posterior marginal distributions of all model random effects and parameters (Martins, Simpson, Lindgren, & Rue, 2013). To select the best candidate models, we evaluated different combinations of spatiotemporal correlation structures and mesh designs, based on cross‐validated predictive ordinate values (Ferkingstad, Held, & Rue, 2017). We created four mesh designs using the constrained refined Delauney triangulation applied to sampling site locations, by varying the sizes of triangles within and outside the sampled area (Figure ), attempting to minimize any boundary effects (Lindgren & Rue, 2015).…”

Section: Methodsmentioning

confidence: 99%

River dams and the stability of bird communities: A hierarchical Bayesian analysis in a tropical hydroelectric power plant

Abreu

Berg

Faria

et al. 2020

Journal of Applied Ecology

View full text Add to dashboard Cite

The effects of anthropogenic disturbance upon the stability of wildlife communities depend on the heterogeneity and connectivity of habitat remnants on multiple scales. The number of hydroelectric dams in biodiversity hotspots (Africa, South America and Asia) is growing rapidly. To establish their environmental impact, it is essential to understand the dynamics of wildlife communities before and following the establishment of dams. We evaluated the impacts of the filling of the Serra do Facão hydroelectric reservoir in the São Marcos river, central Brazil, on the bird community. Using data from 1,145 surveys across 20 sampling sites over 8 years, 2 years before and 6 years after the filling of the reservoir, we assessed the resistance, i.e. maintenance close to an equilibrium state during the disturbance, and resilience, i.e. ability to return to the original state following the disturbance, of the bird community. We used spatiotemporal hierarchical Bayesian models to assess the effects of reservoir filling on five community parameters: abundance, richness, phylogenetic diversity, functional diversity and species composition. In the period subsequent to reservoir filling, there was (a) a marked reduction in bird abundance, richness, phylogenetic diversity and functional diversity, and (b) a reduction in the proportion of forest species, coupled with an increase in the proportion of savanna species. Except for bird abundance, none of the other community attributes returned to their original levels, even after 6 years. Our findings indicate that Cerrado bird communities have both low resistance and low resilience to habitat loss associated with the establishment of hydroelectric reservoirs. Synthesis and applications. The environmental costs of hydroelectric dams are still underestimated or neglected in Brazil. A new paradigm in the assessment of their environmental impacts is warranted, incorporating (a) models of spatiotemporal variations based on long‐term monitoring with surveys initiated before disturbances and (b) measures of functional and phylogenetic diversity, such that society can understand the costs and benefits of the establishment of new hydroelectric dams and make informed decisions. Biodiversity loss could be minimized by ensuring the preservation and connectivity of alluvial habitats, capable of maintaining the supply of resources and the functional and phylogenetic attributes of bird communities associated with such habitats.

show abstract

Fast and accurate Bayesian model criticism and conflict diagnostics using R‐INLA

Cited by 13 publications

References 34 publications

Analysing Multiple Epidemic Data Sources

Analysing Multiple Epidemic Data Sources

Savannah trees buffer herbaceous plant biomass against wild and domestic herbivores

River dams and the stability of bird communities: A hierarchical Bayesian analysis in a tropical hydroelectric power plant

Contact Info

Product

Resources

About