D-Separation
Definition
A graphical criterion that determines whether two sets of nodes in a Bayesian network are conditionally independent given a third set. If nodes A and B are d-separated by set C, observing C renders A and B independent. Used to identify which evidence items remain informative given intermediate findings.
- Field
- Bayesian network graph theory
- Function
- Tests conditional independence between node sets
- Forensic use
- Identifies which evidence items stay informative
- Outcome
- A, B independent when d-separated by set C
Common questions
Why does d-separation matter when a forensic case combines several pieces of evidence in one network?+
It tells the analyst whether learning one piece of evidence makes another piece redundant given what is already known, which prevents double-counting correlated evidence when computing a combined likelihood ratio. Without checking d-separation, an analyst risks treating two dependent pieces of evidence as independent and overstating the evidential weight.
Is d-separation something a forensic report author computes manually?+
No, it is a structural property of the network graph and is normally read off automatically by Bayesian network software once the causal structure between hypotheses and evidence nodes is specified. The analyst's job is to build the network correctly; the software determines the resulting independence relationships.
Related terms
- Belief Propagation
- An algorithm for computing marginal and posterior probabilities in a Bayesian network by passing messages between neighbouring nodes. Exact on tree-structured networks;...
- Conditional Probability Table (CPT)
- A table that specifies the probability distribution of a node given every combination of states of its parent nodes. Every non-root node...
- Directed Acyclic Graph (DAG)
- A graph in which edges have a direction (from parent to child node) and no path can return to a node it...
- Mixture Likelihood Ratio
- The ratio of the probability of observing a mixed DNA profile if the person of interest is a contributor to the probability...
- Sensitivity Analysis
- A technique for assessing how much the posterior probabilities in a Bayesian network change when the values in the conditional probability tables...