[
next
] [
prev
] [
prev-tail
] [
tail
] [
up
]
List of Figures
2.1
Directed Acyclic Graph
2.2
Univariate response: forward and inverse
2.3
Results for various model parameters
3.1
Examples of easy, medium and hard for given new data
3.2
Examples of easy, medium and hard; general new data
3.3
Disjoint-Decomposable Model
3.4
% outside 95% HPD region against number of surfaces modelled
3.5
Non-Disjoint-Decomposable Models
3.6
Error as a Function of Interaction
3.7
Inverse Predictive Distributions: Zero-Inflated Counts
3.8
The Simplex Space
3.9
Nested Compositions: Example
3.10
Nested Compositions: Example
3.11
Nested Compositions: Simplest Case
3.12
Prior Nested Dirichlets
3.13
Posterior Nested Dirichlets
3.14
% outside 95% HPD region; repititions and density, high overdispersion
3.15
% outside 95% HPD region; repititions and density, low overdispersion
4.1
Quadratic Approximation to Poisson
4.2
Gaussian Univariate Approximation
4.3
Quadratic Approximation to Zero-Inflated Poisson
4.4
MCMC and GMRF Approximations: Zero-Inflated Poisson Likelihood
4.5
MCMC Chain Trace Plot: Log-rate
4.6
MCMC Chain Trace Plot: Probability of Presence
4.7
Percentage of Points Outside 95% HPD: Zero-Inflation
5.1
% outside 95% HPD region versus replications
5.2
% outside 95% HPD region versus independent surfaces
5.3
Convergence of inverse predictive distribution
5.4
% outside 95% HPD region; repititions and density
6.1
Modern climates; buffer zone
6.2
Localised Histograms of Proportions Data
6.3
Positive Abundances versus Probability of Presence
6.4
Effect of using the modal hyperparameters
6.5
Distribution of the Δ statistic across taxa; Pollen dataset
6.6
Plot of the Δ statistic against number of taxa
T
6.7
Plot of the
D
statistic against number of taxa
T
6.8
Distribution of the
D
statistic; with and without Gaussian Blurring
6.9
Structure of the Nesting of pollen dataset
6.10
Distribution of
D
l
j