Machine learning of document templates for data extraction
View Patent ↗The present system can perform machine learning of prototypical descriptions of data elements for extraction from machine-readable documents. Document templates are created from sets of training documents that can be used to extract data from form documents, such as: fill-in forms used for taxes; flex-form documents having many variants, such as bills of lading or insurance notifications; and some context-form documents having a description or graphic indicator in proximity to a data element. In response to training documents, the system performs an inductive reasoning process to generalize a document template so that the location of data elements can be predicted for the training examples. The automatically generated document template can then be used to extract data elements from a wide variety of form documents.
1. A method for predicting the placement of a data element within a document, comprising:
using a computer to perform the following steps:
finding an origin of a data element within a document using a data element description;
determining the size of the data element; and
editing the data element.
2. The method of claim 1 , wherein editing the data element includes removing a prefix or a suffix.
3. The method of claim 1 , wherein the data element description includes rules for normalizing data elements and editing the found data element includes normalizing the found data element as specified by the data element description.
4. The method of claim 1 , wherein finding an origin of a data element includes searching for a context phrase that would appear before a data element or after a data element.
5. The method of claim 1 , wherein finding an origin of a data element includes: searching for a context phrase that would appear vertically before or after a data element.
6. The method of claim 1 , wherein finding an origin of a data element includes:
searching for a context phrase that would appear before a data element;
searching for a context phrase that would appear after a data element; and
searching for a context phrase that would appear vertically before or after a data element.
7. The method of claim 1 , wherein the data element description includes one or more alternative data element descriptions.
8. The method of claim 7 , wherein finding an origin of a data element includes finding an origin of a data element for each of the one or more alternative data element descriptions.
9. The method of claim 8 , wherein finding an origin of a data element further includes selecting the best of the one or more alternative data element descriptions.
10. A computer system for predicting the placement of a data element within a document, comprising:
means for finding an origin of a data element within a document using a data element description;
means for determining the size of the data element; and
means for editing the data element.
11. The computer system of claim 10 , wherein the means for editing the data element includes means for removing a prefix or a suffix.
12. The computer system of claim 10 , wherein the data element description includes rules for normalizing data elements and the means for editing the found data element includes means for normalizing the found data element as specified by the data element description.
13. The computer system of claim 10 , wherein the means for finding an origin of a data element includes means for searching for a context phrase that would appear before a data element or after a data element.
14. The computer system of claim 10 , wherein the means for finding an origin of a data element includes means for searching for a context phrase that would appear vertically before or after a data element.
15. The computer system of claim 10 , wherein the means for finding an origin of a data element includes means for:
searching for a context phrase that would appear before a data element;
searching for a context phrase that would appear after a data element; and
searching for a context phrase that would appear vertically before or after a data element.
16. The computer system of claim 10 , wherein the data element description includes one or more alternative data element descriptions.
17. The computer system of claim 16 , wherein the means for finding an origin of a data element includes means for finding an origin of a data element for each of the one or more alternative data element descriptions.
18. The computer systems of claim 17 , wherein the means for finding an origin of a data element further includes means for selecting the best of the one or more alternative data element descriptions.
19. A system for predicting the placement of a data element within a document, comprising:
an optical character recognition module receiving an image representative of a document, the optical character recognition module creating a textual-based representation of a document from a graphical-based representation of a document;
a document template associated with the document type for extracting data from documents of the document type;
an inductive template generator generalizing the document template based on one or more training documents utilizing machine learning; and
a data extraction component utilizing the document template for finding an origin of a data element within a document using a data element description, determining the size of the data element; and editing the data element.
20. The system of claim 19 , wherein editing the data element includes removing a prefix or a suffix.