Method and system for abstracting information for use in link analysis

A method of improving the efficiency of computational link analysis performed by a computer system, said computer system comprising one or more processors and computer-usable non-transitory storage media operationally coupled to the one or more processors, comprising:

• storing a link-oriented data set in the computer-usable non-transitory storage media, wherein the link-oriented data set comprises collected observable data that has been organized into a plurality of typed nodes, each typed node having one or more node values comprising a node type, and a plurality of typed edges between the typed nodes in the link-oriented data set, said plurality of typed edges comprising a plurality of observable data edge-types;
• using a computational process performed by the one or more processors executing software instructions stored in the computer-usable non-transitory storage media, adding a first synthesized edge having a new derived edge-type between a first pair of typed nodes wherein link derivation has been used to define the new derived edge-type, and a second synthesized edge having a new induced edge-type between a second pair of typed nodes, wherein link induction has been used to define the new induced edge-type, wherein the new derived edge-type and the new induced edge-type are not observable data edge-types,
• wherein using link derivation to define a new edge-type comprises storing, in the computer-usable non-transitory storage media, software instructions that comprise the new edge-type and specify one or more dataset operations comprising one or more other edge-types, and wherein using link induction to define a new edge-type comprises storing, in the computer-usable non-transitory storage media, software instructions that comprise the new edge-type and specify one or more functions operating on one or more node values;
• whereby the synthesized edges reveal associations that are not expressed in the collected observable data;
• and
• storing the link-oriented data set with the synthesized edges in the computer-usable non-transitory storage media.

A computer system for improving the efficiency of computational link analysis, comprising:

• One or more processors;
• Computer-usable non-transitory storage media coupled to the one or more processors;
• collected observable data organized as a plurality of typed nodes and a plurality of typed edges between the typed nodes in a link-oriented data set stored in the computer-usable non-transitory storage media, said plurality of typed edges comprising a plurality of observable data edge-types;
• and
• software instructions stored in the computer-usable non-transitory storage media which when executed by a processor cause the computer system to add a first synthesized edge having a new derived edge-type between a first pair of nodes, the new derived edge-type defined by link derivation, and to add a second synthesized edge having a new induced edge-type between a second pair of nodes, the new induced edge-type defined by link induction, the nodes in the first and second pair of nodes comprised in the plurality of typed nodes,
• wherein defining a new edge-type by link derivation comprises storing, in the computer-usable non-transitory storage media, software instructions that comprise the new edge-type and specify one or more dataset operations comprising two or more other edge-types, and defining a new edge-type by link induction comprises storing, in the computer-usable non-transitory storage media, software instructions that comprise the new edge-type and specify one or more functions operating on one or more node values,
• wherein the new derived edge-type and new induced edge-type are not an observable data edge-type, and wherein the first and second synthesized edges reveal one or more associations that are not expressed in the collected observable data.