Collaborative predictive model building

• US 8,160,977 B2
• Filed: 12/07/2007
• Issued: 04/17/2012
• Est. Priority Date: 12/11/2006
• Status: Active Grant

• Current Assignee: Christian D. Poulin
• Original Assignee: Christian D. Poulin
• Inventors: Poulin, Christian D.
• Primary Examiner(s): Gaffin, Jeffrey A
• Assistant Examiner(s): BROWN JR, NATHAN H
• Application Number:
  US11/952,205
• Publication Number:
  US 20080154821A1
• Time in Patent Office: 1,593 Days
• Field of Search: 706/21, 707/706
• US Class Current: 706/21
• 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: G06Q 10/0631 Resource planning, allocation, distributing or scheduling for enterprises or organisations
• CPC Class Codes: G06Q 10/0637 Strategic management or analysis, e.g. setting a goal or target of an organisation; Planning actions based on goals; Analysis or evaluation of effectiveness of goals

Abstract

Techniques for estimating a probability that a future event will occur based on user input includes decomposing a data input stream to build a database of precursor data to build at least one predictive model, building a database of at least one model generated by a model building process using the precursor data in the database, with the at least one model being a model that produces predictions of the likelihood of an event occurring in the future based on analysis of data in the database and storing the at least one model in a second database that stores models, with the database being searchable, to permit the models in the database to be accessed by users. Also disclosed are techniques in which by using a search engine to search a database of models to find a model and a user can query a found model to develop an inference of the likelihood of a future event.

Claims

1. A computer implemented method for estimating a probability that a future event will occur based on user input, the method comprises:

decomposing by one or more computer systems a data input stream to build a first database of precursor data to build at least one predictive model;

building by the one or more computer systems at least one model generated by a model building process using the precursor data in the precursor database, with the at least one model being a model that produces predictions of the likelihood of an event occurring in the future; and

storing by the one or more computer systems the at least one model in a second database that stores models, with the second database being searchable to permit the at least one model in the second database to be accessed by users; and

calculating by the one or more computer systems accuracy of the at least one model against historical data.

2. The method of claim 1 further comprising:

testing by the one or more computer systems the model to optimize the model to refine accuracy level of the model.

3. The method of claim 1 further comprising:

publishing the model.

4. The method of claim 1 wherein building the database comprises:

retrieving data as data strings from a data source;

producing a dataset from the retrieved data strings.

5. The method of claim 4 wherein building the model comprises:

building a statistical model of parent-child relationships from data strings in the dataset by;

determining incidence values for the data strings in the dataset; and

concatenating the incident values with the data strings to provide child variables;

analyzing the child variables and the parent variables to produce statistical relationships between the child variables and a parent variable;

determining probabilities values based on the determined parent child relationships.

6. The method of claim 1 wherein the modeling is based on a pattern recognition technique that can be a Bayesian algorithm, Markov rule-based, neural network or genetic algorithm.

7. The method of claim 1 further comprising:

calculating the accuracy of the model using a k-fold cross-validation.

8. The method of claim 2 wherein analyzing the child variables and the parent variables to produce statistical relationships uses a Bayesian probability algorithm engine.

9. The method of claim 1 further comprising

selecting access permissions for publishing the model.

10. The method of claim 1 further comprising

searching the database for model type.

11. The method of claim 9 further comprising:

performing an inference operation on the model to provide the probability of an event occurring.

12. The method of claim 1 further comprising:

filtering noise from the data retrieved from the data source to provide the data strings.

13. A method executed in a computer system, the method comprises:

using by one or more computer systems a search engine to search a database of models to find a model;

querying by one or more computer systems the model to develop an inference of the likelihood of a future event; and

allowing by the one or more computers a user to discard variables in the model below a certain threshold of relevancy for at least one of calculation time and predictive accuracy.

14. The method of claim 13 further comprising:

providing a market for use of the models.

15. The method of claim 13 wherein in the query a user specifies a parameter of data collection, with at least one of the parameters being a time interval.

16. The method of claim 13 wherein once a model has been published other users can query the model for the probability that a future event will occur.

17. The method of claim 15 wherein an output is an inference or other analysis such as value of information analysis and so forth, that is related to the probability of a model predicting that the future event will occur.

18. The method of claim 13 wherein the building of predictive models relates to health risk factors prediction.

19. The method of claim 13 wherein the building of predictive models relates to genetic traits prediction.

20. The method of claim 13 wherein the building of predictive models relates to commodity price prediction.

21. The method of claim 13 wherein the building of predictive models relates to movie success/revenue prediction.

22. The method of claim 13 wherein the building of predictive models relates to music prediction.

23. The method of claim 13 wherein the building of predictive models relates to political behavior prediction.

24. The method of claim 13 wherein the building of predictive models relates to media bias prediction.

25. The method of claim 13 wherein the building of predictive models relates to games.

26. A computer program product tangibly embodied in a computer readable storage device, the computer program product for estimating a probability that a future event will occur based on user input, the computer program product comprises instructions for causing a computer to:

decompose a data input stream to build a database of precursor data to build at least one predictive model;

build at least one model generated by a model building process using the precursor data in the database, with the at least one model being a model that produces predictions of the likelihood of an event occurring in the future;

build a statistical model of parent-child relationships from data strings in the dataset by instructions to;

determine incidence values for the data strings in the dataset;

concatenate the incident values with the data strings to provide child variables;

analyze the child variables and the parent variables to produce statistical relationships between the child variables and a parent variable;

determine probabilities values based on the determined parent child relationships; and

store the at least one model in a second database that stores models, with the database being searchable to permit the at least one model in the second database to be accessed by users.

27. The computer program product of claim 26, further comprising instructions to:

test the model to optimize the model to refine accuracy level of the model.

28. The computer program product of claim 26, further comprising instructions to:

publish the model.

29. The computer program product of claim 26 wherein instructions to build the database comprises instructions to:

retrieve data as data strings from a data source;

produce a dataset from the retrieved data strings.

30. A computer program product tangibly embodied in a computer readable medium for estimating a probability that a future event will occur based on user input, the computer program product comprises instructions for causing a computer to:

use a search engine to search a database of models to find a model; and

query the model to develop an inference of the likelihood of a future event; and

allow a user to discard variables in the model below a certain threshold of relevancy for at least one of calculation time and predictive accuracy.

31. The computer program product of claim 30 wherein in the query a user specifies a parameter of data collection, with at least one of the parameters being a time interval.