Methods and systems of retrieving documents

• US 9,159,584 B2
• Filed: 05/22/2012
• Issued: 10/13/2015
• Est. Priority Date: 08/18/2000
• Status: Expired due to Fees

• Current Assignee: Hyland Switzerland, Sarl (HSI Holdings II Incorporated)
• Original Assignee: BDGB Enterprise Software SARL (Ninestar Corporation)
• Inventors: Lapir, Gannady, Urbshat, Harry
• Primary Examiner(s): LEROUX, ETIENNE PIERRE
• Application Number:
  US13/477,645
• Publication Number:
  US 20120296933A1
• Time in Patent Office: 1,239 Days
• Field of Search: 707/2, 707/3, 707/5, 707/769, 706/45, 713/165, 382/224, 711/108
• US Class Current: 1/1
• CPC Class Codes: H01L 21/3221 of silicon bodies, e.g. for gettering
• CPC Class Codes: H01L 21/3225 Thermally inducing defects using oxygen present in the silicon body for intrinsic gettering H01L21/3226 takes precedence
• CPC Class Codes: H01L 21/565 Moulds
• CPC Class Codes: H01L 2224/16225 the item being non-metallic, e.g. insulating substrate with or without metallisation
• CPC Class Codes: H01L 23/26 including materials for absorbing or reacting with moisture or other undesired substances , e.g. getters
• CPC Class Codes: H01L 23/3128 the substrate having spherical bumps for external connection
• CPC Class Codes: H01L 23/564 Details not otherwise provided for, e.g. protection against moisture getters H01L23/26
• CPC Class Codes: H01L 27/14698 Post-treatment for the devices, e.g. annealing, impurity-gettering, shor-circuit elimination, recrystallisation
• CPC Class Codes: H01L 2924/15311 being a ball array, e.g. BGA
• CPC Class Codes: H01L 31/186 Particular post-treatment for the devices, e.g. annealing, impurity gettering, short-circuit elimination, recrystallisation
• CPC Class Codes: H10K 50/846 comprising getter material or desiccants
• CPC Class Codes: Y10T 428/28 Web or sheet containing structurally defined element or component and having an adhesive outermost layer

Abstract

Methods and Systems of Retrieving Documents. Coding a plurality of stored documents as respective document feature vectors in a bit-attribute matric, each feature vector comprising an attribute or class. Generating a query vector based on a query document. Performing one or more logical operations between the query vector and the respective document features vectors in the bit-attribute matrix to obtain respective similarity measures. Assigning an attribute or class to the query document based on the attribute or classes of one or more of the feature vectors.

Claims

1. A method of retrieving documents, comprising:

performing processing associated with coding, with a processor, a plurality of stored documents as respective document feature vectors in a bit-attribute matrix, each feature vector comprising at least one attribute or class;

performing processing associated with generating, with the processor, a query vector based on a query document;

performing processing associated with performing, with the processor, one or more logical operations between the query vector and the respective document feature vectors in the bit-attribute matrix to obtain respective similarity measures; and

performing processing associated with assigning, with the processor, an attribute or class to the query document based on the attribute or classes of one or more of the feature vectors.

2. The method of claim 1, further comprising:

performing processing associated with storing, with the processor, each of the respective document feature vectors in respective rows of a bit-attribute matrix.

3. The method of claim 2, wherein the performing processing associated with performing comprises performing one or more logical operations between the query vector and respective rows of the bit-attribute matrix to obtain the respective similarity measures.

4. The method of claim 1, wherein a document feature vector comprises a trigram.

5. The method of claim 1, wherein the query vector comprises a trigram.

6. The method of claim 1, wherein the performing comprises performing one or more logical operations between a trigram of the query feature vector and a trigram of the respective document feature vector to obtain the respective similarity measures.

7. The method of claim 1, further comprising:

performing processing associated with reducing, with the processor, each of the respective document feature vectors into a respective bit string comprising less data than the respective document feature vector; and

performing processing associated with performing, with the processor, one or more logical operations between the query feature vectors and the respective bit strings to obtain the respective similarity measures.

8. The method of claim 1, wherein the respective similarity measure represents how often a feature or a set of features defined by logical 1'"'"'s in the query vector occurs within a corresponding feature vector.

9. The method of claim 1, wherein there is a plurality of result columns which code the respective similarity measures based on how often a feature or a set of features defined by logical 1'"'"'s in the query vector occur within a row of the bit-attribute matrix.

10. The method of claim 1, wherein based on the respective similarity measures a set of candidate documents is chosen for further inspection, the set of candidate documents being further evaluated with respect to their similarity to the query documents being further evaluated with respect to their similarity measures reflecting the similarity between query and candidate under one or more aspects.

11. The method of claim 10, wherein a final similarity measure is obtained based on which documents are finally to be retrieved.

12. The method of claim 10, wherein the further inspection comprises comparing query elements with elements of the candidate documents to obtain one or more of the following further similarity measures:

a measure for a similarity between the textual query elements and textual document elements;

obtaining for each query element a corresponding textual document element based on a similarity measure;

obtaining a measure for the degree of coincidence between a sequential order of the textual query elements and a sequential order of the corresponding textual document elements;

obtaining a measure for a similarity between a distance between elements in the query and a corresponding distance between elements in the document.

13. The method of claim 1, wherein results are returned which lie above a certain similarity threshold, and/or wherein a user definable number of documents are retrieved which have the highest similarity.

14. The method of claim 1, wherein a desired degree similarity can be defined by a user to be:

identity;

an settable degree of similarity or a similarity threshold; and

or identity for certain features and a settable degree of similarity for the other features.

15. The method of claim 1, wherein features coded by a feature vector for their presence or absence comprise one or more of the following;

unigrams, diagrams, multigrams, words, fragments of text, sentences, paragraphs, or any combination thereof;

semantic classes reflecting a certain meaning which elements in the text to be represented may have;

concept classes reflecting an abstract concept to which elements in the text to be represented belong;

a classification class classifying the document according to a classification scheme;

an attribute which may be assigned to a document.

16. The method of claim 15, wherein the classification classes or the attributes may be attributes of the documents, comprising:

date, author, title, publisher, editor, topic of the text, or classification results from classifications by the user or by an automatic classification engine, or any combination thereof.

17. The method of claim 1, wherein the matrix formed by the feature vectors representing the documents to be searched is stored column-wise, and wherein said method further comprises:

for those columns of the matrix where the query feature vector bit is set, performing a processor supported bitwise operation between the columns of the matrix to obtain one or more result columns, the values in the result columns representing the similarity between the query and the stored documents.

18. The method of claim 1, wherein there is a plurality of result columns such that the result columns form a matrix, and wherein a row of the thus formed result column matrix represents a degree of similarity between the query and the stored document corresponding to said row.

19. The method of claim 1 further comprising:

obtaining a set of examples for a certain classification class by carrying out a search for documents belonging to a particular class; and

training a classification engine by said obtained set of examples as the learning input.

20. The method of claim 19, wherein the search for example documents is carried out in a subset of the set of stored documents, the subset being defined through the presence of one or more attributes or classification classes specified by the user in the query, thereby selecting for said query only the documents of said subset.

21. The method of claim 19, the method being applied to realize a self-improving system through the steps of retrieving relevant documents from a set of documents through an associative search;

using the retrieved documents from a set of documents through an associative search;

using the retrieved documents for improving a classification engine, using the classification through a classification engine to improve the retrieval through an associative memory; and

using the improved retrieval to improve the classification engine.