METHOD AND SYSTEM TO PREDICT A DATA VALUE

• US 20110106743A1
• Filed: 01/09/2011
• Published: 05/05/2011
• Est. Priority Date: 01/14/2008
• Status: Active Grant

• Current Assignee: Aptima Inc.
• Original Assignee: Aptima Inc.
• Inventors: DUCHON, ANDREW P.
• Granted Patent: US 8,781,989 B2
• Time in Patent Office: Days
• US Class Current: 706/46
• CPC Class Codes: G06F 16/35 Clustering; Classification

Abstract

Embodiments of the present invention include methods and systems for predicting the likelihood of topics appearing in a set of data such as text. A number of latent variable methods are used to convert the data into a set of topics, topic values and topic profiles. A number of time-course methods are used to model how topic values change given previous topic profiles, or to find historical times with similar topic values and then projecting the topic profile forward from that historical time to predict the likelihood of the topics appearing. Embodiments include utilizing focus topics, such as valence topics, and data representing financial measures to predict the likelihood of topics. Methods and systems for modeling data and predicting the likelihood of topics over other dimensions are also contemplated.

Claims

1. A computer based system for predicting a data value, said system comprising:

a processor capable of executing machine instructions; and

the machine instructions including a means for predicting a predicted data value of a first data set at a predicted dimension value utilizing a focus topic profile.

2. The computer based system of claim 1 further comprising:

a means for analyzing a second set of data using a latent variable method using at least one focus topic; and

the focus topic profile comprises at least one focus topic value of the at least one focus topic from the second set of data.

3. The computer based system of claim 1 further comprising:

a means for analyzing a second set of data using a latent variable method;

the latent variable method using at least one focus feature from a third data set to create at least one focus topic from the second data set; and

the focus topic profile comprises at least one focus topic value of the at least one focus topic from the second set of data.

4. The computer based system of claim 1 wherein the focus topic is a valence topic.

5. The computer based system of claim 1 wherein the means for predicting the predicted data value comprises using a predictive model of a temporal relationship of the focus topic profile and the first data set.

6. The computer based system of claim 1 wherein the means for predicting the predicted data value comprises using a time series analysis model of a temporal relationship of the focus topic profile and the first data set.

7. The computer based system of claim 1 wherein the means for predicting the predicted data value further comprises a modeling controller component capable of:

receiving a base dimension value having a base focus topic value;

receiving a predicted dimension value having a difference dimension value from the base dimension value;

analyzing the focus topic profile over at least one dimension value to identify a most similar focus topic value to the base focus topic value;

the dimension value at the most similar focus topic value being a selected dimension value; and

selecting at least one focus topic value from the focus topic profile at a difference dimension value from the selected dimension value as the predicted data value at the predicted dimension value.

8. A computer based system for predicting a data value, said system comprising:

a processor capable of executing machine instructions on a first set of data comprising financial data; and

the machine instructions including a means for predicting a predicted data value of the first data set at a predicted dimension value utilizing a focus topic profile.

9. The computer based system of claim 8 wherein the financial data comprises a function of a Dow Jones Industrial Average.

10. The computer based system of claim 8 wherein the focus topic profile comprises a focus topic profile of news data.

11. The computer based system of claim 8 wherein the focus topic profile comprises a valence topic profile.

12. The computer based system of claim 8 further comprising:

a means for analyzing a second set of data using a latent variable method to create at least one focus topic; and

the focus topic profile comprises at least one focus topic value of the at least one focus topic.

13. The computer based system of claim 8 further comprising:

a means for analyzing a second set of data using a latent variable method;

the latent variable method using at least one focus feature from a third data set to create at least one focus topic from the second data set; and

the focus topic profile comprises at least one focus topic value of the at least one focus topic from the second set of data.

14. The computer based system of claim 8 wherein the means for predicting the predicted data value comprises using a predictive model of a temporal relationship of the focus topic profile and the first data set.

15. The computer based system of claim 8 wherein the means for predicting the predicted data value comprises using a time series analysis model of a temporal relationship of the focus topic profile and the first data set.

16. The computer based system of claim 8 wherein the means for predicting the predicted data value further comprises a modeling controller component capable of:

receiving a base dimension value having a base focus topic value;

receiving a predicted dimension value having a difference dimension value from the base dimension value;

analyzing the focus topic profile over at least one dimension value to identify a most similar focus topic value to the base focus topic value;

the dimension value at the most similar focus topic value being a selected dimension value; and

selecting at least one focus topic value from the focus topic profile at a difference dimension value from the selected dimension value as the predicted data value at the predicted dimension value.

17. A computer based system for predicting a data value, said system comprising:

a processor capable of executing machine instructions on a first set of data comprising financial data; and

the machine instructions including a means for transforming a second set of data over at least one dimension value to create a topic profile and a means for predicting a predicted data value of the first set of data at a predicted dimension value.

18. The computer based system of claim 17 wherein the financial data comprises a function of a Dow Jones Industrial Average.

19. The computer based system of claim 17 wherein:

the means for transforming the second set of data further comprises analyzing the second set of data using a latent variable method to create at least one topic; and

the topic profile comprises at least one topic value of the at least one topic.

20. The computer based system of claim 17 wherein:

the means for transforming the second set of data further comprises analyzing the second set of data using latent Dirichlet allocation to create at least one topic; and

the topic profile comprises at least one topic value of the at least one topic.

21. The computer based system of claim 17 wherein the means for predicting the predicted data value of the first set of data comprises using a predictive model of the topic profile.

22. The computer based system of claim 17 wherein the means for predicting the predicted data value of the first set of data comprises using a time series analysis model of a temporal relationship of the topic profile and the first data set.

23. The computer based system of claim 17 wherein the means for predicting the predicted data value of the first set of data further comprises a modeling controller component capable of:

receiving a base dimension value having a base topic value;

receiving the predicted dimension value having a difference dimension value from the base dimension value;

analyzing the topic profile over at least one dimension value to identify a most similar topic value to the base topic value;

the dimension value at the most similar topic value being a selected dimension value; and

selecting at least one topic value from the topic profile at a difference dimension value from the selected dimension value as the predicted data value at the predicted dimension value.

24. A method for predicting a data value, said method comprising;

receiving a first data set; and

predicting a predicted data value of the first data set at a predicted dimension value utilizing a focus topic profile.

25. The method of claim 24 further comprising:

analyzing a second data set using a latent variable method using at least one focus topic; and

the focus topic profile comprises at least one focus topic value of the at least one focus topic from the second data set.

26. The method of claim 24 further comprising:

analyzing a second set of data using a latent variable method;

the latent variable method using at least one focus feature from a third data set to create at least one focus topic from the second data set; and

the focus topic profile comprises at least one focus topic value of the at least one focus topic from the second set of data.

27. The method of claim 24 wherein the focus topic is a valence topic.

28. The method of claim 24 wherein the step of predicting the predicted data value comprises using a predictive model of a temporal relationship of the focus topic profile and the first data set.

29. The method of claim 24 wherein the step of predicting the predicted data value comprises using a time series analysis model of a temporal relationship of the focus topic profile and the first data set.

30. The method of claim 24 wherein the step of predicting the predicted data value comprises:

receiving a base dimension value having a base focus topic value;

receiving a predicted dimension value having a difference dimension value from the base dimension value;

analyzing the focus topic profile over at least one dimension value to identify a most similar focus topic value to the base focus topic value;

the dimension value at the most similar focus topic value being a selected dimension value; and

selecting at least one focus topic value from the focus topic profile at a difference dimension value from the selected dimension value as the predicted data value at the predicted dimension value.