CONTENT AND QUALITY ASSESSMENT METHOD AND APPARATUS FOR QUALITY SEARCHING

• US 20090077068A1
• Filed: 08/20/2008
• Published: 03/19/2009
• Est. Priority Date: 05/14/2004
• Status: Active Grant

• Current Assignee: Constantin Aliferis, Yin Aphinyanaphongs
• Original Assignee: Constantin Aliferis, Yin Aphinyanaphongs
• Inventors: Aphinyanaphongs, Yin, Aliferis, Constantin
• Granted Patent: US 8,275,772 B2
• Time in Patent Office: Days
• US Class Current: 1/1
• CPC Class Codes: G06F 16/355 Class or cluster creation or modification
• CPC Class Codes: G16H 70/60 relating to pathologies

Abstract

A computer-based process retrieves information organized in documents containing text and/or coded representations of text. The process involves obtaining and labeling a selected set of documents based on content quality, and extracting and representing features from each document in the selected set. The extracted and selected features are modified, and models are constructed using parametric learning algorithms. The constructed models are capable of assigning a label to each document. The model parameters are instantiated using a first subset of the selected set of documents. Parameters are chosen by validating the corresponding model against at least a second subset of the full document set. The constructed models also are capable of assigning labels to similar documents outside a selected subset not previously given to the process of model construction.

Claims

1. A computer-based process for filtering information according to content quality that is organized in documents comprising the steps of:

obtaining a selected set of documents;

labeling each document of the selected set based on content quality;

extracting and representing features from each document in the selected set;

modifying the extracted and represented features;

constructing models using a pattern recognition algorithm, the constructed models being capable of assigning a label based on the content quality of each document, model parameters being instantiated using a first subset of the selected set of documents, and the parameters being chosen by validating the corresponding model against at least a second subset of the selected set, wherein during a search for related documents, the validated model labels the documents resulting from the search based on content quality.

2. A process according to claim 1, further comprising displaying the label of at least one of documents resulting from the search.

3. A process according to claim 2, wherein all negatively labeled documents are presented in the search results.

4. A process according to claim 2, wherein negatively labeled documents are not presented in the search results.

5. A process according to claim 1, further comprising applying the validated models to at least a third subset of the initial document set mutually exclusive of the first and second subsets to evaluate performance.

6. A process according to claim 1, wherein the step of validating the models is performed iteratively.

7. A process according to claim 5, wherein the steps of validating the corresponding model against at least a second subset of the selected set and applying the validated models to at least the third subset of the initial document set mutually exclusive of the first and second subsets is performed iteratively to evaluate performance.

8. A process according to claim 1, wherein the set of documents are chosen using a plurality of different sources.

9. A process according to claim 1, wherein the step of extracting and representing features from each document in the selected set is performed using a plurality of different methods.

10. A process according to claim 1, wherein the step of modifying the extracted and represented features from each document in the selected set is performed using a plurality of different methods.

11. A process according to claim 1, wherein the set of documents is comprised of web pages.

12. A process of using a pattern recognition algorithm to filter documents based on quality, comprising the steps of:

obtaining a first set of documents;

obtaining a model for labeling documents based on content quality, constructed by the pattern recognition algorithm using a second set of documents that is related to the first set of documents;

using the model to label the first set of documents, wherein the label of at least one of the documents of the first set is displayed.

13. A process according to claim 12 wherein a model is obtained before the first set of documents is obtained.

14. A process according to claim 12, wherein all the labels of negatively labeled documents of the first set are displayed.

15. A process according to claim 12, wherein negatively labeled documents are removed from the first set of documents.

16. A process according to claim 12, wherein before obtaining the first set of documents, the first set of documents had been previously labeled by the model.

17. A process according to claim 12, wherein obtaining the first set of documents is done by a search algorithm.

18. A process according to claim 12, wherein the first and second set of documents are comprised of web pages.

19. A system using a model to filter documents according to quality comprising:

a storage device;

at least one processor programmed to;

obtain a first set of documents labeled according to content quality from the storage device;

extract and represent features from the first set of document;

modify the extracted and represented features;

construct models for labeling documents based on content quality using a pattern recognition algorithm stored in the storage device;

use a first subset of the first set of documents to instantiate parameters for the models;

use a second subset of the first set of documents to validate the models to select the parameters of the model;

obtain a second set of documents, related to the first set of documents;

use the validated model to label the second set of documents according to content quality;

distinguish between labels of the second set of documents.

20. A system according to claim 19, wherein the first and second set of documents are web pages.

21. A system according to claim 19, further comprising a server wherein the storage device is located on the server.

22. A system according to claim 19, wherein the at least one processor is programmed to iteratively use the first subset of the first set of documents to instantiate parameters for the models.

23. A system according to claim 19, wherein the at least one processor is programmed to iteratively use a second subset of the first set of documents to validate the models to select the parameters of the models.

24. A system according to claim 19, wherein the at least one processor is programmed to apply the validated models to a third subset of the first set of documents, that is mutually exclusive of the first and second subsets to evaluate performance.

25. A system according to claim 24, wherein the at least one processor is programmed to iteratively apply the validated models to a third subset of the first set of documents, that is mutually exclusive of the first and second subsets to evaluate performance.