Method of detecting and responding to changes in the online community's interests in real time

• US 20100094879A1
• Filed: 10/14/2008
• Published: 04/15/2010
• Est. Priority Date: 03/30/2007
• Status: Active Grant

• Current Assignee: Needlebot, Inc.
• Original Assignee: Can Deniz Akyuz, Stuart Donnelly
• Inventors: Akyuz, Can Deniz, Donnelly, Stuart
• Granted Patent: US 8,275,773 B2
• Time in Patent Office: Days
• US Class Current: 707/749
• CPC Class Codes: G06F 16/93 Document management systems
• CPC Class Codes: G06F 16/9535 Search customisation based on user profiles and personalisation

Abstract

A method of locating relevant documents wherein documents are given a fingerprint comprising weights associated with particular topic categories of a classification system, each weight representing a degree to which the document relates to the particular topic category. Documents whose fingerprints have a predetermined degree of mathematical overlap with the fingerprint may be considered relevant. A user is alerted to new relevant documents. Advertisers can offer advertisements near search results that achieve a predefined amount of relevance to text submitted by the advertiser rather than bidding on keywords. Unwanted content may be blocked from the search and filters to further refine the search may be used. E-mail spam may be blocked using textual relevance rather than keywords. Visual cues linked to a hierarchy of relevance help display the relevant documents. The methods may be used in combination with keyword searching.

Claims

1. A method of alerting a user that new documents relevant to the user'"'"'s search have entered a database, comprising:

calculating a reference fingerprint as a reference point;

classifying new documents that enter the database so that individual new documents have a fingerprint, the fingerprint comprising a list of weights associated with particular topic categories in a classification system, each of the weights representing a degree to which the document relates to the particular topic category that the weight is associated with, the weights obtained automatically from a computer program;

comparing the fingerprint for each classified new document with the reference fingerprint, and

alerting the user about new documents whose fingerprints have a threshold amount of mathematical overlap with the reference fingerprint.

2. The method of claim 1, wherein the reference fingerprint is a user fingerprint, the user fingerprint comprising a list of cumulative weights associated with particular topic categories in a classification system, each of the cumulative weights representing a degree to which text or texts recently accessed by a user relates to the particular topic category that the cumulative weight is associated with, the cumulative weights obtained from weights that in turn were obtained automatically from a computer program.

3. The method of claim 1, wherein the reference fingerprint is a target fingerprint, the target fingerprint representing a fingerprint from a target piece of text submitted by a user and comprising a list of weights associated with particular topic categories in a classification system, each of the weights in the target fingerprint representing a degree to which the target piece of text relates to the particular topic category that the weight in the fingerprint for the target piece of text is associated with, the weights in the target fingerprint for the target piece of text obtained automatically from a computer program.

4. The method of claim 1, further comprising setting a level of precision for calculating the mathematical overlap, the level of precision represented by a number of digits contained in classification codes associated with the particular topic categories.

5. The method of claim 1, wherein in classifying new documents that enter the database only those new documents that have certain keywords are classified.

6. The method of claim 1, further comprising notifying the user of documents relevant to the user'"'"'s search that are already in the database by comparing the reference fingerprint to a fingerprint for each document in an inverted list of documents in the database, the inverted list comprising for each topic category of the classification system a weight associated with a particular document of the list of documents, the weight representing a degree to which the particular document relates to said each topic category, the weights obtained automatically from a computer program and alerting the user about those documents whose fingerprints have a threshold amount of mathematical overlap with the reference fingerprint.

7. A method for presenting advertisements to a user for display on a user computing device, wherein the advertisements are selected from a plurality of advertisements, the method comprising:

obtaining user fingerprints of users who search using a search engine;

obtaining from an advertiser an advertisement fingerprint, the advertising fingerprint masking weights for particular topic categories designated by the advertiser;

computing a mathematical overlap between the advertisement fingerprint and particular user fingerprints of the users; and

displaying the advertisement to users whose user fingerprints reach a threshold mathematical overlap.

8. The method of claim 7, further comprising displaying the advertisement to users who are also searching using search requests that contain particular keywords.

9. A method of re-ranking a list of documents obtained from a search, the fingerprints of the documents in the list of documents being known, the method comprising:

identifying a piece of unwanted text;

creating a fingerprint for the piece of unwanted text, the fingerprint comprising a list of weights associated with particular topic categories in a classification system, each of the weights in the fingerprint for the piece of unwanted text representing a degree to which the piece of unwanted text relates to the particular topic category that the weight in the fingerprint for the piece of unwanted text is associated with, the weights in the fingerprint for the piece of unwanted text obtained automatically from a computer program;

searching the list of documents by comparing the fingerprint for the piece of unwanted text with the fingerprint for each document in the list of documents;

re-ranking the list of documents based on a degree to which a document in the list has a mathematical overlap with the fingerprint of the piece of unwanted text.

10. The method claim 9, further comprising re-ranking the list of documents so that documents in the list are given a higher ranking based on a degree to which such document has a lower mathematical overlap with the fingerprint of the piece of unwanted text.

11. The method of claim 9, further comprising re-ranking the list of documents by removing from the list those documents whose fingerprints have a threshold amount of mathematical overlap with the fingerprint of the piece of unwanted text.

12. A method of locating relevant documents within a universe of documents, the documents of said universe having been classified so that each document in the universe has a fingerprint, the fingerprint comprising a list of weights associated with particular topic categories in a classification system, each of the weights representing a degree to which the document relates to the particular topic category that the weight is associated with, the weights obtained automatically from a computer program, the method comprising:

identifying a piece of unwanted text;

creating a fingerprint for the piece of unwanted text, the fingerprint comprising a list of weights associated with particular topic categories in the classification system, each of the weights in the fingerprint for the piece of unwanted text representing a degree to which the piece of unwanted text relates to the particular topic category that the weight in the fingerprint for the piece of unwanted text is associated with, the weights in the fingerprint for the piece of unwanted text obtained automatically from a computer program;

creating an anti-fingerprint from the fingerprint by calculating, for each of the weights in the fingerprint, a mathematical inverse of that weight and using that mathematical inverse in place of that weight;

a sum of a particular weight and its mathematical inverse equaling a maximum value for the particular weight;

searching all or a portion of the universe of documents by comparing the anti-fingerprint with the fingerprint for each document in that all or a portion of the universe of documents; and

ranking the documents whose fingerprints have a predetermined degree of mathematical overlap with the fingerprint of the piece of unwanted text.

13. The method of claim 12, wherein the vectors that the fingerprints represent are normalized and a sum of a particular weight and its mathematical inverse equals one.

14. The method of claim 12, further comprising setting a level of precision for the anti-fingerprint, the level of precision represented by a number of digits contained in classification codes associated with the particular topic categories.

15. A method of re-ranking a list of documents obtained during a search, the documents in the list having been ranked based on a mathematical overlap between a fingerprint of each the documents in the list of documents with a reference fingerprint, the method comprising:

identifying a piece of unwanted text;

creating a fingerprint for the piece of unwanted text, the fingerprint comprising a list of weights associated with particular topic categories in the classification system, each of the weights in the fingerprint for the piece of unwanted text representing a degree to which the piece of unwanted text relates to the particular topic category that the weight in the fingerprint for the piece of unwanted text is associated with, the weights in the fingerprint for the piece of unwanted text obtained automatically from a computer program;

creating an anti-fingerprint from the fingerprint by calculating, for each of the weights in the fingerprint, a mathematical inverse of that weight and using that mathematical inverse in place of that weight;

a sum of a particular weight and its mathematical inverse equaling a maximum value for the particular weight;

creating a revised fingerprint for the documents in the list of documents by subtracting the anti-fingerprint from the fingerprint of each document in the list; and

re-ranking the documents in the list of documents.

16. A method of re-ranking a list of documents obtained during a search, comprising:

calculating a fingerprint for each document in the list, the fingerprint comprising a list of weights associated with particular topic categories in a classification system, each of the weights in the fingerprint for the document representing a degree to which the document relates to the particular topic category that the weight in the fingerprint for the document is associated with, the weights in the fingerprints for the documents obtained automatically from a computer program;

ranking the documents whose fingerprints have a predetermined degree of mathematical overlap with a reference fingerprint;

identifying a piece of unwanted text;

creating a fingerprint for the piece of unwanted text, the fingerprint comprising a list of weights associated with particular topic categories in the classification system, each of the weights in the fingerprint for the piece of unwanted text representing a degree to which the piece of unwanted text relates to the particular topic category that the weight in the fingerprint for the piece of unwanted text is associated with, the weights in the fingerprint for the piece of unwanted text obtained automatically from a computer program;

creating an anti-fingerprint from the fingerprint by calculating, for each of the weights in the fingerprint, a mathematical inverse of that weight and using that mathematical inverse in place of that weight;

a sum of a particular weight and its mathematical inverse equaling a maximum value for the particular weight;

creating a revised fingerprint for the documents in the list of documents by subtracting the anti-fingerprint from the fingerprint of each document in the list; and

re-ranking the documents in the list of documents.

17. A method of filtering e-mail containing text, comprising:

identifying a piece of unwanted text;

creating a fingerprint for the piece of unwanted text, the fingerprint comprising a list of weights associated with particular topic categories in a classification system, each of the weights in the fingerprint for the piece of unwanted text representing a degree to which the piece of unwanted text relates to the particular topic category that the weight in the fingerprint for the piece of unwanted text is associated with, the weights in the fingerprint for the piece of unwanted text obtained automatically from a computer program;

calculating a fingerprint of the text in the e-mail, the fingerprint comprising a list of weights associated with particular topic categories in the classification system, each of the weights in the fingerprint for the text in the e-mail representing a degree to which the text in the e-mail relates to the particular topic category that the weight is associated with, the weights in the fingerprint for the text in the e-mail obtained automatically from a computer program;

calculating a degree of mathematical overlap between the fingerprint of the piece of unwanted text and the fingerprint of the text of the e-mail; and

blocking a particular e-mail when the degree of mathematical overlap between the fingerprint of the piece of unwanted text and the fingerprint of the text of the particular e-mail reached a threshold amount.

18. A method of displaying results of a search for textual content based on relevance, the method comprising:

obtaining from the search a list of documents each of which has associated with it a measure of relevance;

defining a hierarchy of visual cues;

assigning visual cues to the documents in the list based on a linkage of a hierarchy of the relevance of the documents in the list to the hierarchy of the visual cues; and

displaying the list of documents having the visual cues.

19. The method of claim 18, wherein the visual cues include, alone or in combination, color, font, style of text, presence of icons associated with particular text, presence of background color and presence foreground color.