System and method for predictive analysis and predictive analysis markup language

• US 7,389,282 B2
• Filed: 11/02/2005
• Issued: 06/17/2008
• Est. Priority Date: 11/02/2004
• Status: Expired due to Fees

• Current Assignee: Viziant Corp.
• Original Assignee: Viziant Corp.
• Inventors: Maren, Alianna J.
• Primary Examiner(s): STARKS, WILBERT L
• Application Number:
  US11/264,748
• Publication Number:
  US 20060167689A1
• Time in Patent Office: 958 Days
• Field of Search: 706/45, 706/52, 703/2
• US Class Current: 706/52
• CPC Class Codes: G06N 5/01 Dynamic search techniques; Heuristics; Dynamic trees; Branch-and-bound
• CPC Class Codes: G06Q 10/04 Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem" market predictions or forecasting for commercial activities G06Q30/0202
• CPC Class Codes: G10L 25/12 the extracted parameters being prediction coefficients

Abstract

A predictive architecture system and method that supports predicting likely future evolutions by aggregating and analyzing large amounts of data. The invented system includes a multi-level architecture, along with one or more feedback functions that drive search and discovery for information in support of a “hypothesis” or discovery of significant information in advance of clear hypothesis formation and passing on higher-level predictive support or lack of support for existing and/or new hypotheses.

Claims

1. A system for predictive analysis, comprising:

a predictive analysis processor configured to receive a data input and to generate a plurality of weighted course-of-action predictions, wherein the predictive analysis processor is operatively coupled to a modeler for defining a contextual world model in which future scenarios may be predicted;

a predictive analysis feedback processor, operatively coupled to the first predictive analysis processor and configured to receive the plurality of weighted course-of action predictions and automatically providing supporting information to the predictive analysis processor based on the application of a nonlinear gradient to the received weighted course-of-action predictions; and

an output device configured to display the course of action predictions.

2. The system of claim 1, further comprising:

a model and simulation processor configured to project events, actions and relationships carried out by an entity as defined in the contextual world model;

wherein the predictive analysis processor further comprises a utility function for determining projected events of greatest value to the entity initiating the projected events;

a scenario generation processor configured to simulate execution of each projected event, action and relationship and aggregating the utility function to each projected event, action and relationship; and

a synthesis processor configured to identify efficient events, actions and relationships and creating a generalized function based on the simulated execution of each projected event, action and relationship.

3. The system of claim 2, wherein the model and simulation processor is configured to construct a taxonomy of future events such that there are multiple levels of granularity in the representation of these events.

4. The system of claim 1, wherein the modeler is configured to define any one of a number of contextual models including but not limited to stakeholder interaction, geopolitical interaction, economical interaction, climate interaction, geospatial interaction and environmental interaction.

5. The system of claim 1, wherein the predictive analysis feedback processor is configured to apply emergent behavior information as the nonlinear gradient.

6. The system of claim 1, wherein the predictive analysis feedback processor is configured to apply diverse scenario integration information as the nonlinear gradient.

7. The system of claim 1, wherein the predictive analysis feedback processor is configured to apply one or more accelerants and/or forcing functions as the nonlinear gradient.

8. A system for predictive analysis, comprising:

a knowledge discovery processor configured to collect and transforming a set of structured data and/or semi-structured data and/or unstructured data elements into a filtered set of information;

a markup language modeling processor configured to define a markup language consistent with a subject entity'"'"'s ontology, wherein the markup language is applied to the filtered set of information;

a hypothesis generator, operatively coupled to the knowledge discovery processor, configured to characterize the filtered set of information, wherein a first feedback controller is operatively coupled between the hypothesis generator and the knowledge discovery processor for automatically providing supporting information to the knowledge discovery processor based on the characterized filtered set of information;

a predictive analysis processor, operatively coupled to the hypothesis generator, for generating a plurality of weighted course-of action predictions, wherein a second feedback controller is operatively coupled between the first predictive analysis processor and the hypothesis generator for automatically providing supporting information to the hypothesis generator based on the weighted course-of action predictions;

a predictive analysis feedback processor, operatively coupled to the first predictive analysis processor, for receiving the plurality of weighted course-of action predictions and automatically providing supporting information to the predictive analysis processor based on the application of a nonlinear gradient to the received weighted course-of-action predictions.

9. The system of claim 8, further comprising:

a first model processor configured to produce a contextual model in which the weighted course of action predictions are generated; and

a second model processor configured to provide input to the first model processor in the form of a plurality of world models.

10. The system of claim 9, wherein the first model processor is configured to define any one of a number of contextual models including but not limited to stakeholder interaction, geopolitical interaction, economical interaction, climate interaction, geospatial interaction and environmental interaction.

11. The system of claim 8, wherein the knowledge discovery processor is configured to produce a filtered set of information about the activities of a plurality of entities and/or information about the communication between entities.

12. The system of claim 8, wherein the knowledge discovery processor is configured to produce a filtered set of information capable of being used to characterize a global environment that will impact an entity.

13. The system of claim 8, further comprising:

a model and simulation processor configured to project events, actions and relationships carried out by an entity as defined in the contextual world model;

wherein the predictive analysis processor further comprises a utility function for determining projected events of greatest value to the entity initiating the projected events;

a scenario generation processor configured to simulate execution of each projected event, action and relationship and aggregating the utility function to each projected event, action and relationship; and

a synthesis processor configured to learn efficient events, actions and relationships generating a creating a generalized function based on the simulated execution of each projected event, action and relationship.

14. The system of claim 8, wherein the predictive analysis feedback processor is configured to apply emergent behavior information as the nonlinear gradient.

15. The system of claim 8, the predictive analysis feedback processor is configured to apply diverse scenario integration information as the nonlinear gradient.

16. The system of claim 8, wherein the predictive analysis feedback processor is configured to apply one or more accelerants and/or forcing functions as the nonlinear gradient.

17. The method of claim 16, wherein the generating step further comprises:

projecting events, actions and relationships carried out by an entity as defined in the contextual world model;

determining projected events of greatest value to the entity initiating the projected events;

simulating execution of each projected event, action and relationship and aggregating the utility function to each projected event, action and relationship; and

learning efficient events, actions and relationships and creating a generalized learned function based on the simulated execution of each projected event, action and relationship.

18. The method of claim 16, wherein the nonlinear gradient is emergent behavior information.

19. The method of claim 16, wherein the nonlinear gradient is diverse scenario integration information.

20. The method of claim 16, wherein the nonlinear gradient is one or more accelerants and/or forcing functions.

21. A method for conducting predictive analysis, comprising the steps of:

providing a data input to a processor;

defining a contextual world model in which future scenarios may be predicted;

generating a plurality of weighted course-of-action predictions;

applying a non-linear gradient to the plurality of weighted course-of action predictions to generate supporting information; and

regenerating a plurality of weighted course-of-action predictions using the supporting information;

displaying the predictions on an output device.