Analytic systems, methods, and computer-readable media for structured, semi-structured, and unstructured documents

1. A non-transitory computer-readable medium storing instructions that, when executed by a computer, cause the computer to execute steps comprising:

accessing, from one or more computer readable storage devices, a compilation of one or more electronically stored documents;

receiving information regarding a defined value, wherein the defined value has a defined format, has a defined two-dimensional spatial relationship to words on a selected page, and is associated with one or more keywords;

for a page of the compilation,

identifying words and contender values on the page using optical character recognition (OCR) and post-OCR processing, and

grouping the identified words and the identified contender values into anchor blocks based on spatial positioning on the page;

for each identified contender value,

numerically determining a first confidence that the contender value is associated with the defined value based on a comparison of a calculated two-dimensional spatial relationship between the contender value and the anchor blocks on the page with the defined two-dimensional spatial relationship,

numerically determining a second confidence that the contender value is associated with the defined value based on a comparison of words in the anchor blocks on the page with the one or more keywords, and

numerically determining a third confidence that the contender value is associated with the defined value based at least in part on a comparison of a format of the contender value with the defined format; and

identifying the contender value as positively associated with the defined value based at least in part on the first confidence, the second confidence, and the third confidence; and

for each of the anchor blocks comprising a contender value on the page,

assigning a first anchor block weight to all words in the anchor block,

assigning a second anchor block weight to all words in a second anchor block above the anchor block on the page,

assigning a third anchor block weight to all words in a third anchor block to the left of the anchor block on the page, and

assigning a fourth anchor block weight, lower than the first anchor block weight, the second anchor block weight, and the third anchor block weight, to a remaining anchor block, based on a two-dimensional spatial distance to the anchor block,

wherein the second anchor block is immediately adjacent to the anchor block such that there are no anchor blocks between the second anchor block and the anchor block in the vertical direction.

2. The non-transitory computer-readable medium of claim 1 , wherein assigning the fourth anchor block weight comprises assigning a lower anchor block weight when the fourth anchor block is located to the right of the anchor block than when the fourth anchor block is located an equivalent two-dimensional spatial distance to the left of the anchor block.

3. The non-transitory computer-readable medium of claim 1 , wherein the determination of the first confidence is adjusted based on the number of anchor blocks on the page.

4. The non-transitory computer-readable medium of claim 1 , wherein the determination of the third confidence is based at least in part on a compensation for typographical differences between words in the anchor blocks on the page and the words on the selected page not exceeding a numerical threshold.

5. The non-transitory computer-readable medium of claim 1 , wherein the determination of the third confidence is based at least in part on a compensation for root-stem associations between words in the anchor blocks on the page and the words on the selected page.

6. The non-transitory computer-readable medium of claim 1 , wherein the determination of the third confidence is based at least in part on relative positions of words in the anchor blocks on the page.

7. The non-transitory computer-readable medium of claim 1 , wherein grouping the identified words and the identified contender values into anchor blocks is further based on typographical characteristics of the identified words and the identified contender values, the typographical characteristics comprising font size, font face, or a combination thereof.

8. The non-transitory computer-readable medium of claim 1 , wherein the determination of the third confidence is based at least in part on a compensation for typographical differences between the contender value and the defined value not exceeding a numerical threshold.

9. The non-transitory computer-readable medium of claim 1 , wherein the post-OCR processing is configured to identify a contender value based on formatting including numerical formatting, date formatting, delimiting character formatting, or a combination thereof.

10. The non-transitory computer-readable medium of claim 1 , wherein the format of the defined value comprises a location of the defined value in a zone on the selected page, and wherein the determination of the third confidence is based at least in part on whether the contender value is located in the zone.

11. A non-transitory computer-readable medium storing instructions that, when executed by a computer, cause the computer to execute steps comprising:

accessing, from one or more computer readable storage devices, a compilation of one or more electronically stored documents;

receiving information regarding a defined value, wherein the defined value has a defined format, has a defined two-dimensional spatial relationship to words on a selected page, and is associated with one or more keywords;

for a page of the compilation,

identifying words and contender values on the page using optical character recognition (OCR) and post-OCR processing, and

grouping the identified words and the identified contender values into anchor blocks based on spatial positioning on the page;

for each identified contender value,

numerically determining a first confidence that the contender value is associated with the defined value based on a comparison of a calculated two-dimensional spatial relationship between the contender value and the anchor blocks on the page with the defined two-dimensional spatial relationship,

numerically determining a second confidence that the contender value is associated with the defined value based on a comparison of words in the anchor blocks on the page with the one or more keywords, and

numerically determining a third confidence that the contender value is associated with the defined value based at least in part on a comparison of a format of the contender value with the defined format; and

identifying the contender value as positively associated with the defined value based at least in part on the first confidence, the second confidence, and the third confidence; and

for each of the anchor blocks comprising a contender value on the page,

assigning a first anchor block weight to all words in the anchor block,

assigning a second anchor block weight to all words in a second anchor block above the anchor block on the page,

assigning a third anchor block weight to all words in a third anchor block to the left of the anchor block on the page, and

assigning a fourth anchor block weight, lower than the first anchor block weight, the second anchor block weight, and the third anchor block weight, to a remaining anchor block, based on a two-dimensional spatial distance to the anchor block,

wherein the third anchor block is immediately adjacent to the anchor block such that there are no anchor blocks between the third anchor block and the anchor block in the horizontal direction.

12. The non-transitory computer-readable medium of claim 11 , wherein the determination of the third confidence is based at least in part on an order of the page within the compilation.

13. The non-transitory computer-readable medium of claim 11 , wherein the electronically stored documents further comprise a structured document and the determination of the third confidence is based at least in part on the contender value's position in a defined location on the page.

14. The non-transitory computer-readable medium of claim 11 , wherein, when the defined value is a ZIP code, the determination of the third confidence is based at least in part on an evaluation of a position of the contender value within its associated anchor block.

15. The non-transitory computer-readable medium of claim 11 , wherein, when the defined value is a ZIP code, the determination of the third confidence is based at least in part on a comparison of the contender value with a compendium of valid ZIP codes.

16. The non-transitory computer-readable medium of claim 11 , wherein, when the defined value is an address or a portion thereof, the determination of the third confidence is based at least in part on a validation of the contender value against a network-accessible geolocation library.

17. The non-transitory computer-readable medium of claim 11 , wherein the first confidence is associated with a first dimension, the second confidence is associated with a second dimension, and the third confidence is associated with a third dimension, and the steps further comprise adjusting a distinct weight assigned to each of the first dimension, the second dimension, and the third dimension based on the defined value when extracting positive contender values as positively associated with the defined value.

18. The non-transitory computer-readable medium of claim 11 , wherein the determination of the third confidence is based at least in part on a compensation for typographical differences between words in the anchor blocks on the page and the words on the selected page not exceeding a numerical threshold.

19. The non-transitory computer-readable medium of claim 11 , wherein the determination of the third confidence is based at least in part on a compensation for root-stem associations between words in the anchor blocks on the page and the words on the selected page.

20. The non-transitory computer-readable medium of claim 11 , wherein the determination of the third confidence is based at least in part on relative positions of words in the anchor blocks on the page.