Bayesian networks (BNs)
q Use probability theory for representing uncertainty
q Represents a probability distribution graphically (directed acyclic graphs)
q Nodes: random variables (discrete, continuous)
q Arcs indicate conditional dependencies between v

Learning Bayesian Networks
Linear and Discrete Models
Learning Network Parameters
Linear Coefficients
Learning Probability Tables
Learning Causal Structure
Conditional Independence Learning
Statistical Equivalence
TETRAD II
Bayesian Learning of Bayesi

Parameter Estimation
Data variables: initially given uniform distribution
Sparse or no data variables: elicited.
Experts were asked to report confidence in estimates (equivalent sample size
ESS), to be used by data learning/training method: EM (Expectati

Expert Evaluation
Test aspects of network not represented in data set
Conditions required for healthy native fish communities
Robustness of network
Fish Ecologists
Environmental Managers / Natural Resource Managers
GBC Bayesian Networks
Risk Management fr

Structure terminology andlayout
Family metaphor:
Parent Child
Ancestor Descendant
Markov Blanket = parents + children + childrensparent
s Tree analogy:
root node: no parents
leaf node: no children
intermediate node: non-leaf, non-root
Layout conventi

Parameterisation
Expert Elicitation used to parameterise aspects of model not represented by
data (lack of variability in data set)
Iterative process of updating expert derived probabilities (prior probabilities)
using data (automated process)
Needed t

Bayesian Networks
Address Uncertainty and Complexity
Increasingly being used in ecological applications
Modular DSS
Complex system composed of simpler parts (or multiple models)
Inputs: expert opinion, literature, data, other models
Predictions able

Types of Reasoning
Ecological Risk Assessment
Process for improving environmental management to protect ECOLOGICAL
values/assets
Focus: Managing physical, chemical and biological processes to protect
ecological endpoints
Ecological sustainability/integ

Experiment 1: ROC Results
Experiment 2: ROC Results
Summary of Results
Experiment I (Models of whole data)
q PROCAM model does at least as well as Busselton
On Busselton data
For both "relative" (ROC) and "absolute" (BIR) risk
q CaMML Models do as well
Bu

Bayesian Networks
Data Structure which represents the dependence between variables.
Gives concise specification of the joint probability distribution.
A Bayesian Network is a graph in which thefollowing holds:
1. A set of random variables makes up the nod