Automated data entry system

• US 7,596,545 B1
• Filed: 08/29/2005
• Issued: 09/29/2009
• Est. Priority Date: 08/27/2004
• Status: Expired due to Fees

• Current Assignee: University of Kansas
• Original Assignee: University of Kansas
• Inventors: Perry, Steven M., Lee, Danico, Tsatsoulis, Constantinos
• Primary Examiner(s): Fleurantin; Jean B
• Application Number:
  US11/214,212
• Time in Patent Office: 1,492 Days
• Field of Search: 707 1- 10, 707100-1041, 707200-206, 715221-226
• US Class Current: 1/1
• CPC Class Codes: G06F 40/274 Converting codes to words; Guess-ahead of partial word inputs
• CPC Class Codes: G06Q 30/02 Marketing; Price estimation or determination; Fundraising
• CPC Class Codes: G06Q 30/0603 Catalogue ordering

Abstract

The present invention is a novel system and method for providing automated data entry of information. The system may include a processing subsystem capable of processing extensible markup language documents, a predicting subsystem capable of reviewing entered data of an input document and providing a suggestion to the input document, and a database for storing parsed extensible markup language documents. Database may maintain a plurality of relationships among received data of previous documents, and predictor may analyze a plurality of relationships for suggestions. The method for providing automated data entry of information may monitor received data, analyze relationships among received data, and predict a data entry field. A data entry field may be based on received data and relationships among received data. Predicting of data entry fields may incorporate a plurality of predictors whereby one or more values from values provided by each of the plurality of predictors is provided to a user.

Claims

1. A system for automatic data entry of information, comprising:

a processor, said processor being configured to receive a plurality of documents and parse data entered in said plurality of documents;

a database structure, said database structure being configured to store parsed data,

wherein a plurality of relationships developed from said parsed data is maintained; and

an ensemble predictor, said ensemble predictor being configured to review entered data of an input document and said plurality of relationships from said received plurality of documents and provide a suggestion for an uncompleted field of said input document said ensemble predictor including;

a plurality of individual predictors; and

a suggestion aggregator, said suggestion aggregator providing a single suggestion based upon a voting scheme among said plurality of individual predictors, at least one of said plurality of individual predictors including a weighting system determining performance of an individual predictor of said plurality of individual predictors and weights votes of an individual predictor based upon past performance.

2. The system as claimed in claim 1, wherein said documents are extensible markup language documents.

3. The system as claimed in claim 1, wherein said processing subsystem parses data of a completed document and stores the parsed data of said completed document in said database.

4. The system as claimed in claim 1, wherein said plurality of relationships developed from said parsed data includes node-value pairs and attribute-value pairs.

5. The system as claimed in claim 1, wherein said plurality of individual predictors include at least one of a statistical predictor, an inductive predictor or an instance-based predictor.

6. The system as claimed in claim 5, wherein said statistical predictor is a naï

ve Bayesian classifier.

7. The system as claimed in claim 5, wherein said inductive predictor is a C4.5 inductive learning predictor.

8. The system as claimed in claim 5, wherein said instance-based predictor is a K-Nearest-Neighbor predictor.

9. The system as claimed in claim 1, wherein said weighting system assigns a weight to each individual predictor of the plurality of individual predictors based on a percentage of correctly provided suggestions of a total number of suggestions.

10. The system as claimed in claim 9, wherein said weighting system adapts to multiple data domain types through adjustment of weighting for each individual predictor of said plurality of individual predictors based on past performance.

11. The system as claimed in claim 1, wherein said suggestion aggregator provides a single suggestion based upon a voting scheme among the plurality of predictors in combination with weighting of each vote from each individual predictor.

12. A method for automating data entry of information, comprising:

monitoring data received from a plurality of documents via a processor configured to receive said plurality of documents and parse data entered in said plurality of documents;

analyzing relationships among said data received from a plurality of documents via an ensemble predictor; and

predicting a data entry field via said ensemble predictor;

determining performance of each individual predictor of said plurality of individual predictors via said ensemble predictor;

weighing votes of an individual predictor based upon past performance via a weighting system of said ensemble predictor, wherein a plurality of predicted data entries are provided based on said data received from a plurality of documents and relationships among said data received from a plurality of documents, and

providing a single suggestion from a plurality of predicted data entries for said data entry field via a suggestion aggregator, wherein said single suggestion is selected from predicted data entries.

13. The method as claimed in claim 12, wherein said single suggestion is selected by the suggestion aggregator based upon a voting scheme among said plurality of predicted data entries.

14. The method as claimed in claim 12, further comprising measuring an error rate of a prediction via said suggestion aggregator.

15. The method as claimed in claim 14, wherein said single suggestion selected by the suggestion aggregator is based upon a voting scheme and said error rates of said prediction.

16. A system for automating data entry of information, comprising:

a processor, said processor being configured to receive a plurality of documents and parse data entered in said plurality of documents;

a database structure, said database structure being configured to store parsed data, wherein a plurality of relationships developed from said parsed data is maintained; and

an ensemble predictor, said ensemble predictor including;

a plurality of individual predictors, said plurality of individual predictors including at least one of a statistical predictor, an inductive predictor or an instance-based predictor, wherein each individual predictor is configured to review entered data of an input document and said plurality of relationships from said received plurality of documents and submit at least one suggestion for a uncompleted field of said input document;

a weighting system configured to analyze performance of each individual predictor and weigh votes of each individual predictor based upon past performance; and

a suggestion aggregator configured to provide one or more data entries from said at least one suggestion from each individual predictor based upon a voting scheme among the plurality of predictors in combination with weighting of each vote from each individual predictor.

17. The system as claimed in claim 16, wherein said documents are extensible markup language documents.

18. The system as claimed in claim 16, wherein said processor parses data of a completed document and stores the parsed data of said completed document in said database.

19. The system as claimed in claim 16, wherein said plurality of relationships developed from said parsed data includes node-value pairs and attribute value pairs.