Risk analysis using unstructured data

• US 9,141,686 B2
• Filed: 11/08/2012
• Issued: 09/22/2015
• Est. Priority Date: 11/08/2012
• Status: Active Grant

• Current Assignee: Bank of America Corp.
• Original Assignee: Bank of America Corp.
• Inventors: Kern, Daniel C., Hogeboom, David A., Bromstead, Anne
• Primary Examiner(s): Reyes, Mariela
• Assistant Examiner(s): VU, THONG H
• Application Number:
  US13/672,012
• Publication Number:
  US 20140129561A1
• Time in Patent Office: 1,048 Days
• Field of Search: 707/101, 707/102, 707/600, 707/719, 707/5, 707/1/1, 707/748, 707/708, 705/2
• US Class Current: 1/1
• CPC Class Codes: G06F 16/285 Clustering or classification
• CPC Class Codes: G06F 16/345 Summarisation for human users

Abstract

Unstructured data is received from a plurality of sources to facilitate risk analysis. The unstructured data comprises a plurality of bodies of text. Each body of text from the unstructured data is deconstructed into individual terms. The individual terms from each body of text are converted into a structured form. The individual terms in the structured form are categorized according to a comparison of the structured form to another structured form. The individual terms in the structured form are quantified according to at least the categorization of the individual terms.

Claims

1. A system, comprising:

a network interface operable to receive unstructured data from a plurality of data sources, the plurality of data sources comprising a competitor database, a vendor database, and a marketing database, wherein the unstructured data relates to a financial risk of an organization and comprises a plurality of text documents, each text document comprising a plurality of groups of words;

a processor communicatively coupled to the network interface and operable to;

deconstruct each group of words from the unstructured data into individual words;

convert the individual words from each group of words into a plurality of structured forms, each structured form corresponding to a single group of words;

determine a numerical value associated with each individual word according to;

a number of times the individual word appears in the group of words and an association of the group of words with a risk experienced by an organization, wherein each structured form is a vector that includes each individual word and the numerical value associated with the individual word;

compare each structured form to another structured form using a Bayesian inference;

categorize the individual words in each structured form into at least one category according to the comparison and the at least one category is selected from a set of categories consisting of organization name, geographical region, organization size, number of employees, number of countries represented, public organization, private organization, regulatory body, industry, and fine amount, the categories indicating the financial risk of the organization; and

quantify the individual words in each structured form according to at least the categorization of the individual words by weighting each individual word.

2. The system of claim 1, wherein the processor is further operable to remove insignificant words from each group of words before converting the individual words into the structured form.

3. The system of claim 1, wherein:

the processor is further operable to create a report based on the categorized and quantified individual words; and

the network interface is further operable to communicate the report to a computer to facilitate analysis of the individual words and the associated group of words.

4. The system of claim 1, wherein the network interface is further operable to receive the unstructured data from at least one of an external data source and an internal data source.

5. The system of claim 1, wherein the processor is further operable to:

determine whether individual words have associated quantifiable data; and

link the individual words to quantifiable data.

6. The system of claim 1, wherein the processor is further operable to:

compare a first structured form to a second structured form;

identify similar text between the first structured form and the second structured form to determine a similar text group; and

determine a size of the similar text group.

7. Non-transitory computer readable medium comprising logic, the logic, when executed by a processor, operable to:

receive unstructured data from a plurality of data sources, the plurality of data sources comprising a competitor database, a vendor database, and a marketing database, wherein the unstructured data relates to a financial risk of an organization and comprises a plurality of text documents, each text document comprising a plurality of groups of words;

deconstruct each group of words from the unstructured data into individual words;

convert the individual words from each group of words into a plurality of structured forms, each structured form corresponding to a single group of words;

determine a numerical value associated with each individual word according to;

a number of times the individual word appears in the group of words and an association of the group of words with a risk experienced by an organization, wherein each structured form is a vector that includes each individual word and the numerical value associated with the individual word;

compare each structured form to another structured form using a Bayesian inference;

categorize the individual words in each structured form into at least one category according to the comparison and the at least one category is selected from a set of categories consisting of organization name, geographical region, organization size, number of employees, number of countries represented, public organization, private organization, regulatory body, industry, and fine amount, the categories indicating the financial risk of the organization; and

quantify the individual words in each structured form according to at least the categorization of the individual words by weighting each individual word.

8. The computer readable medium of claim 7, wherein the logic is further operable to remove insignificant words from each group of words before converting the individual words into the structured form.

9. The computer readable medium of claim 7, wherein the logic is further operable to:

create a report based on the categorized and quantified individual words; and

communicate the report to a computer to facilitate analysis of the individual words and the associated group of words.

10. The computer readable medium of claim 7, wherein the logic is further operable to:

determine whether individual words have associated quantifiable data; and

link the individual words to quantifiable data.

11. The computer readable medium of claim 7, wherein the logic is further operable to:

compare a first structured form to a second structured form;

identify similar text between the first structured form and the second structured form to determine a similar text group; and

determine a size of the similar text group.

12. A method, comprising:

receiving unstructured data from a plurality of data sources, the plurality of data sources comprising a competitor database, a vendor database, and a marketing database, wherein the unstructured data relates to a financial risk of an organization and comprises a plurality of text documents, each text document comprising a plurality of groups of words;

deconstructing, by a processor, each group of words from the unstructured data into individual words;

converting, by the processor, the individual words from each group of words into a plurality of structured forms, each structured form corresponding to a single group of words;

determining, by the processor, a numerical value associated with each individual word according to;

a number of times the individual word appears in the group of words and an association of the group of words with a risk experienced by an organization, wherein each structured form is a vector that includes each individual word and a quantification of the individual word;

comparing, by the processor, each structured form to another structured form using a Bayesian inference;

categorizing, by the processor, the individual words in each structured form into at least one category according to the comparison and the at least one category is selected from a set of categories consisting of organization name, geographical region, organization size, number of employees, number of countries represented, public organization, private organization, regulatory body, industry, and fine amount, the categories indicating the financial risk of the organization; and

quantifying, by the processor, the individual words in each structured form according to at least the categorization of the individual words by weighting each individual word.

13. The method of claim 12, further comprising removing insignificant words from each group of words before converting the individual words into the structured form.

14. The method of claim 12, further comprising:

creating a report based on the categorized and quantified individual words; and

communicating the report to a computer to facilitate analysis of the individual words and the associated group of words.

15. The method of claim 12, wherein receiving the unstructured data from the plurality of sources comprises receiving the unstructured data from at least one of an external data source and an internal data source.

16. The method of claim 12, wherein quantifying the individual words comprises weighting the individual words based on quantifiable data.

17. The method of claim 12, wherein quantifying the individual words comprises:

determining whether individual words have associated quantifiable data; and

linking the individual words to quantifiable data.

18. The method of claim 12, further comprising:

comparing a first structured form to a second structured form;

identifying similar text between the first structured form and the second structured form to determine a similar text group; and

determining a size of the similar text group.

19. The method of claim 12, wherein each individual term in the structured form is associated with a number that indicates the number of times the individual word appears in the group of words.