Cognitive initialization of large-scale advection-diffusion models

1. A computer-implemented method for determining a first numerical value for a first parameter of a computer-based advection-diffusion model, the method comprising:

determining the first parameter for a computer-based advection-diffusion model for which the first numerical value is to be determined;

performing a natural language processing (NLP) search for a plurality of numerical values for the first parameter on a plurality of documents;

selecting the first numerical value for the first parameter from the plurality of numerical values for the first parameter, based on the first numerical value for the parameter appearing more times in the plurality of documents than a second numerical value for the parameter appearing in the plurality of documents;

performing the advection-diffusion model using the first numerical value for the first parameter; and

storing a result of performing the advection-diffusion model in a computer memory.

2. The computer-implemented method of claim 1 , wherein a first document of the plurality of documents is an unstructured document.

3. The computer-implemented method of claim 2 , wherein the unstructured document is a natural-language document.

4. The computer-implemented method of claim 1 , wherein a first document of the plurality of documents comprises a structured document.

5. The computer-implemented method of claim 4 , wherein the structured document comprises a relational database.

6. A computer-implemented method for determining a first numerical value for a first parameter of a computer-based advection-diffusion model, the method comprising:

determining the first parameter for a computer-based advection-diffusion model for which the first numerical value is to be determined;

performing a natural language processing (NLP) search for a plurality of numerical values for the first parameter on a plurality of documents;

selecting the first numerical value for the first parameter from the plurality of numerical values for the first parameter based on the first numerical value for the parameter appearing in a first document of the plurality of documents, there being more citations to the first document than citations to a second document of the plurality of documents, the second document including a second value for the parameter of the plurality of values for the parameter;

performing the advection-diffusion model using the first numerical value for the first parameter; and

storing a result of performing the advection-diffusion model in a computer memory.

7. A computer system for determining a first numerical value for a first parameter of a computer-based advection-diffusion model, comprising:

a processor, a computer-readable memory, a computer-readable tangible storage device, and program instructions stored on the storage device for execution by the processor via the memory, wherein execution of the program instructions by the computer system configures the computer system to:

perform a natural language processing (NLP) search for a plurality of numerical values for the first parameter on a plurality of documents;

select the first numerical value for the first parameter from the plurality of numerical values for the first parameter;

perform the advection-diffusion model using the first numerical value for the first parameter; and

halt performing the NLP search when a predetermined number of numerical values for the first parameter have been found.

8. The computer system of claim 7 , wherein execution of the program instructions further configures the computer system to:

select the first numerical value for the first parameter from the plurality of numerical values for the first parameter based on the first numerical value being associated with a similar geographical location as a geographical location being modeled with the advection-diffusion model.

9. The computer system of claim 7 , wherein execution of the program instructions further configures the computer system to:

select the first numerical value for the first parameter from the plurality of numerical values for the first parameter based on the first numerical value being associated with a similar weather chemistry as a weather chemistry being modeled with the advection-diffusion model.

10. The computer system of claim 7 , wherein execution of the program instructions further configures the computer system to:

select the first numerical value for the first parameter from the plurality of numerical values for the first parameter based on the first numerical value being associated with a similar grid resolution as a grid resolution being modeled with the advection-diffusion model.

11. The computer system of claim 7 , wherein execution of the program instructions further configures the computer system to:

select the first numerical value for the first parameter from the plurality of numerical values for the first parameter based on the first numerical value being associated with a similar climate as a climate being modeled with the advection-diffusion model.

12. The computer system of claim 7 , wherein execution of the program instructions further configures the computer system to:

select the first numerical value for the first parameter from the plurality of numerical values for the first parameter based on the first numerical value being associated with a similar market volatility as a market volatility being modeled with the advection-diffusion model.

13. A computer program product, comprising:

a computer readable storage medium having programming instructions embodied therewith, the program instructions executable by a computer cause the computer to:

perform a natural language processing (NLP) search for a plurality of numerical values for a first parameter of an advection-diffusion model on a plurality of documents;

determine whether the plurality of numerical values are associated with the first parameter based on performing a statistical classification on the plurality of numerical values in the plurality of documents;

select a first numerical value for the first parameter from the plurality of numerical values for the first parameter; and

perform the advection-diffusion model using the first numerical value for the first parameter.

14. The computer program product of claim 13 , wherein the program instructions executable by the computer further cause the computer to:

access the plurality of documents that are stored on a remote computer and accessed via a communications network.

15. The computer program product of claim 13 , wherein the program instructions executable by the computer further cause the computer to:

select the first numerical value for the first parameter from the plurality of numerical values for the first parameter based on the first numerical value for the parameter appearing more times in the plurality of documents than a second numerical value for the parameter appears in the plurality of documents, or based on the first numerical value for the parameter appearing in a first document of the plurality of documents, there being more citations to the first document than citations to a second document of the plurality of documents, the second document including a second value for the parameter of the plurality of values for the parameter.

16. The computer program product of claim 13 , wherein the program instructions executable by the computer further cause the computer to:

store a result of performing the advection-diffusion model; and

using the result of performing the advection-diffusion model to determine a second numerical value for a second parameter of a second advection-diffusion model.

17. The computer program product of claim 13 , wherein the program instructions executable by the computer further cause the computer to:

before performing the NLP search, training a component that performs the search with labeled documents.