DISTENUM

DISTENUM

Description

An enumerated type of distribution types.

Values

Value Description

deCondSparse A "conditional sparse" distribution is like a full distribution in that it has a row for every combination of parent assignments and a column for every state of the node. It is sparse in the sense that some rows can use default parameter values rather than explicit values.

deCondCI A "conditional causally independent" distribution has a row for every abnormal state of a parent and one "leak" row telling the probability when all parents are in their normal state. Like all these distributions, it has one column for every state of the node.

deCondGraph A "conditional graph" distribution is a generalization of a decision tree distribution. It has one row for leaf in the graph. Like all these distributions, it has one column for every state of the node.

Examples

See: Distributions: an Overview

Also See

Add (ModelNodes), Type (Dist)

Value	Description
deCondSparse	A "conditional sparse" distribution is like a full distribution in that it has a row for every combination of parent assignments and a column for every state of the node. It is sparse in the sense that some rows can use default parameter values rather than explicit values.
deCondCI	A "conditional causally independent" distribution has a row for every abnormal state of a parent and one "leak" row telling the probability when all parents are in their normal state. Like all these distributions, it has one column for every state of the node.
deCondGraph	A "conditional graph" distribution is a generalization of a decision tree distribution. It has one row for leaf in the graph. Like all these distributions, it has one column for every state of the node.