Unified numerical and semantic analytics system for decision support

• US 8,538,915 B2
• Filed: 07/12/2010
• Issued: 09/17/2013
• Est. Priority Date: 07/12/2010
• Status: Active Grant

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Drissi, Youssef, Grandison, Tyrone W., Harrison, Colin G., Katircioglu, Kaan K., Paraszczak, Jurij R.
• Primary Examiner(s): Richardson, James E
• Application Number:
  US12/834,615
• Publication Number:
  US 20120011139A1
• Time in Patent Office: 1,163 Days
• Field of Search: 707/603, 707/710, 707/766, 707/768, 707/765
• US Class Current: 707/603
• CPC Class Codes: G06F 16/2452 Query translation

Abstract

A system and method for responding to a query. An original query is received. A first semantic query and a second semantic query are extracted from the information of the original query. The first semantic query is transformed, based upon semantic analysis, into a numeric model query. The second semantic query is submitted to a semantic analyzer and the numeric model query is submitted to a numeric model analyzer. A response for the second semantic query and a response for the numeric model query are integrated into an answer for the original query.

Claims

1. A method for responding to a query, the method comprising:

performing the following with a processor;

receiving an original query for information related to an action affecting a public population;

extracting a first semantic query and a second semantic query from the original query, the first semantic query comprising a quantitative action term;

transforming, based upon semantic analysis of the quantitative action term, the first semantic query into a numeric model query;

determining semantic response data for the second semantic query based upon semantic analysis of unstructured data sources that comprise unstructured data pertaining to behaviors of the public population to determine response information for the original query;

determining numeric response data for the numeric model query based upon at least one mathematical model incorporating numeric data associated with the public population to determine response information for the original query; and

integrating the semantic response data and the numeric response data into an answer for the original query.

2. The method of claim 1, further comprising accepting the original query from a user as a natural language question.

3. The method of claim 1, further comprising:

extracting an additional numeric model query from the original query; and

determining additional numeric response data for the additional numeric model query based upon at least one mathematical model incorporating numeric data associated with the public population,

wherein the integrating further comprising integrating the additional numeric response data into the answer for the original query.

4. The method of claim 1,

wherein the determining numeric response data comprises:

identifying additional data for the at least one mathematical model to use to respond to the numeric model query;

determining the additional data based upon semantic analysis of the unstructured data sources that comprise unstructured data pertaining to behaviors of the public population; and

incorporating the additional data into the at least one mathematical model,

wherein the numeric response data is further based upon the at least one mathematical model incorporating the additional data.

5. The method of claim 1, further comprising:

providing the answer to a user;

determining, subsequent to providing the answer to the user and subsequent to the user taking the action, semantic reaction data based upon semantic analysis of the unstructured data sources that comprise unstructured data pertaining to behaviors of the public population to determine a reaction of the public population to the action; and

updating, based on the semantic reaction data, at least one of the at least one mathematical model and a model used for semantic analysis of the unstructured data sources.

6. The method of claim 5, wherein the unstructured data sources comprise at least one of at least one blog, at least one letter, and at least one data stream describing behaviors of the public population.

7. The method of claim 1, further comprising:

maintaining a repository storing at least data obtained from the unstructured data sources, and

wherein the semantic analysis comprises performing a semantic analysis of the unstructured data stored in the repository.

8. The method of claim 7, wherein the maintaining the repository comprises receiving, from a plurality of sources, unstructured data provided by members of the public population, the unstructured data reflecting behaviors of the public population.

9. A system for responding to a query, the system comprising:

a memory;

a processor communicatively coupled to the memory, wherein the processor comprises;

an instruction processor configured to;

receive an original query for information related to an action affecting a public population;

extract a first semantic query and a second semantic query from the original query, the first semantic query comprising a quantitative action term;

transform, based upon semantic analysis of the quantitative action term, the first semantic query into a numeric model query;

determine semantic response data for the second semantic query based upon semantic analysis of unstructured data sources that comprise unstructured data pertaining to behaviors of the public population to determine response information for the original query;

determine numeric response data for the numeric model query based upon at least one mathematical model incorporating numeric data associated with the public population to determine response information for the original query; and

integrate the semantic response data and the numeric response data into an answer for the original query.

10. The system of claim 9, the instruction processor receives the original query from a user as a natural language question.

11. The system of claim 9, the instruction processor further configured to:

extract an additional numeric model query from the original query; and

determine additional numeric response data for the additional numeric model query based upon at least one mathematical model incorporating numeric data associated with the public population,

wherein the instruction processor is further configured to integrate the additional numeric response data into the answer for the original query.

12. The system of claim 9, wherein the instruction processor is configured to determine numerical response data by being configured to:

identify additional data for the at least one mathematical model to use to respond to the numeric model query;

determine the additional data based upon semantic analysis of the unstructured data sources that comprise unstructured data pertaining to behaviors of the public population; and

incorporate the additional data into the at least one mathematical model,

wherein the numeric response data is further-based upon the additional data.

13. The system of claim 9, wherein the instruction processor is further configured to:

provide the answer to a user;

determine, subsequent to providing the answer to the user and subsequent to the user taking the action, semantic reaction data based upon semantic analysis of the unstructured data sources that comprise unstructured data pertaining to behaviors of the public population to determine a reaction of the public population to the action;

update, based on the semantic reaction data, at least one of the at least one mathematical model and a model used for semantic analysis of the unstructured data sources.

14. The system of claim 13, wherein the unstructured data sources comprise at least one of at least one blog, at least one letter, and at least one data stream describing behaviors of the public population.

15. A computer program product for responding to a query, the computer program product comprising:

a non-transitory storage device readable by a processing circuit and storing instructions for execution by the processing circuit for performing a method comprising;

receiving an original query for information related to an action affecting a public population;

extracting a first semantic query and a second semantic query from the original query, the first semantic query comprising a quantitative action term;

transforming, based upon semantic analysis of the quantitative action term, the first semantic query into a numeric model query;

determining semantic response data for the second semantic query based upon semantic analysis of unstructured data sources that comprise unstructured data pertaining to behaviors of the public population to determine response information for the original query;

determining numeric response data for the numeric model query based upon at least one mathematical model incorporating numeric data associated with the public population to determine response information for the original query; and

integrating the semantic response data and the numeric response data into an answer for the original query.

16. The computer program product of claim 15, the method further comprising receiving the original query from a user as a natural language question.

17. The computer program product of claim 15, the method further comprising:

extracting an additional numeric model query from the original query; and

determining additional numeric response data for the additional numeric model query based upon at least one mathematical model incorporating numeric data associated with the public population,

wherein the integrating further comprising integrating the additional numeric response data into the answer for the original query.

18. The computer program product of claim 15, the method further comprising:

wherein the determining numeric response data comprises;

identifying additional data for the at least one mathematical model to use to respond to the numeric model query;

determining the additional data based upon semantic analysis of the unstructured data sources that comprise unstructured data pertaining to behaviors of the public population; and

incorporating the additional data into the at least one mathematical model,

wherein the numeric response data is further based upon the at least one mathematical model incorporating the additional data.

19. The computer program product of claim 15, the method further comprising:

providing the answer to a user;

determining, subsequent to providing the answer to the user and subsequent to the user taking the action, semantic reaction data based upon semantic analysis of the unstructured data sources that comprise unstructured data pertaining to behaviors of the public population to determine a reaction of the public population to the action; and

updating, based on the semantic reaction data, at least one of the at least one mathematical model and a model used for semantic analysis of the unstructured data sources.

20. The computer program product of claim 19, wherein the unstructured data sources comprise at least one of at least one blog, at least one letter, and at least one data stream describing behaviors of the public population.