Structural Priors

Overview

A Structural Prior assigns prior knowledge to an individual arc within a range of -1.0 to 1.0, where -1.0 strongly increases the cost of adding the arc during structural learning and 1.0 strongly reduces it. This allows prior knowledge to be defined more locally than is possible with Local Structural Coefficients, which operate at the node level. Structural Priors can be saved to and loaded from a Dictionary, a simple text file that can be opened and edited with any text editor.

Dictionary Syntax

The syntax of a Dictionary containing Structural Priors is analogous to that of Dictionaries of other attributes. Each line names the start node, the end node, and the value of the Structural Prior. As an alternative to the equal sign ”=” shown in the example below, you can also use a space or a tab.

FBS->K=0.076695684161
Condition->K=1.0
BP->K=0.076695684161
EF-TTE->FBS=0.068965584386
Condition->FBS=1.0
DM->FBS=1.0
PR->FBS=0.076695684161

Usage

To import a Dictionary of Structural Priors to your current Bayesian network, select Data > Associate Dictionary > Arc > Structural Priors from the Menus.