| 主权项 |
1. A method for analyzing a probabilistic graphical model, comprising:
retrieving a logical graph model, the logical graph model having one or more variables wherein each variable is represented as a logical class with attributes, wherein the logical graph model includes the variables of a probabilistic graphical model; selecting a variable in the logical graph model and comparing the selected variable to other variables in the logical graph model; generating, for one or more variables in the logical graph model, a value subclass for each of the variable's scenarios when the user asserts discrete scenarios for a variable; generating, for each template variable, index subclasses, each index subclass having at least one attribute comprised of a property and an object, the latter a subclass of the index object class; asserting at least one scenario or factor for a variable; creating edges with labels between the variables of the logical graph model, wherein the labeled edges correspond to influence relationships in the probabilistic graphical model, wherein the labels include “influences” and “influenced by” inverse properties; and creating additional index subclasses for one or more variables of the logical graph model, wherein the additional index subclasses of the one or more variables are influenced by additional index subclasses of one or more other variables in the probabilistic graphical model. |
