Dynamic search engine and database

• US 20030046311A1
• Filed: 06/19/2002
• Published: 03/06/2003
• Est. Priority Date: 06/19/2001
• Status: Abandoned Application

• Current Assignee: Biozak Incorporated
• Original Assignee: Biozak Incorporated
• Inventors: Baidya, Ryan, Miftakhov, Valery
• Application Number:
  US10/177,346
• Publication Number:
  US 20030046311A1
• Time in Patent Office: Days
• US Class Current: 707/200
• CPC Class Codes: G06F 16/954 Navigation, e.g. using categorised browsing

Abstract

An industry database and method of creating same is provided. The database is created in accordance with a process that includes: identifying a plurality of web sites meeting at least one search criteria; automatically extracting URL addresses for each of the plurality of web sites; automatically categorizing each of the web sites and their corresponding URL addresses in accordance with a predefined category structure; and automatically indexing and storing each of the URL addresses in accordance with the predefined category structure in the database. A method of using a database system is also provided. The method includes: storing in a database, information extracted from a plurality of web sites, wherein the information is automatically categorized and indexed in accordance with a predefined category structure and includes a plurality of URL addresses corresponding to the plurality of web sites; receiving a user query; executing a search engine in response to the user query that searches a subset of the stored information extracted from a subset of the plurality of web sites, and subsequently searching said subset of web sites to find additional information responsive to said user query.

Claims

1. A method of creating an industry database, comprising:

conducting an Internet search for information meeting at least one search criteria;

creating a first list of URL addresses corresponding to web pages identified as a result of said Internet search;

unstemming said URL addresses in said first list to create a second list of URL addresses corresponding to unique web sites;

comparing said second list of URL addresses to URL addresses previously stored in said database;

deleting URL addresses from said second list that are duplicative of URL addresses previously stored in said database so as to create a third list of URL addresses;

automatically categorizing at least one URL address from said third list as belonging to a predefined category; and

automatically indexing and storing said at least one URL under said predefined category in said database.

2. The method of claim 1 wherein said step of automatically categorizing comprises:

selecting a subset of URL addresses from said third list so as to specify a training set for creating a statistical model;

downloading content from web sites corresponding to said subset of URL addresses;

creating a first word count list for each web site corresponding to said subset of URL addresses;

manually discarding at least one word determined to be a non-discriminating word from said first word count lists, thereby creating a second word count list for each of said web sites;

manually classifying each URL address from said subset as either belonging to said predefined category or not belonging to said predefined category based on said content from said web sites corresponding to the subset of URL addresses;

creating a statistical model representative of word count characteristics exhibited by web sites belonging to said predefined category and those web sites not belonging to said predefined category, based on said second word count lists;

validating said statistical model on said training set of web sites;

automatically downloading content from a web site corresponding to said at least one URL address from said third list; and

automatically comparing said content from said web site corresponding to said at least one URL address to said statistical model so as to automatically categorize said at least one URL as either belonging to or not belonging to said predefined category.

3. The method of claim 2 further comprising calculating a confidence score based on said step of automatically comparing said content to said statistical model, wherein if said confidence score is below a threshold value, said at least one URL is presented to a human administrator for review.

4. The method of claim 2 wherein said statistical model further represents site structure characteristics of said web sites corresponding to said subset of URL addresses.

5. The method of claim 1 further comprising automatically extracting at least one company name associated with said at least one URL and, thereafter, automatically indexing and storing said at least one company name under said predefined category in said database.

6. The method of claim 5 wherein said step of automatically extracting said at least one company name comprises:

identifying and counting word phrase frequencies from web site content associated with said at least one URL, thereby creating a first list of word phrase frequencies;

identifying and counting word phrase frequencies in content from a plurality of web sites associated with URL addresses in said second or third lists of URL addresses, thereby creating a second list of word phrase frequencies; and

comparing said first list of word phrase frequencies with said second list of word phrase frequencies to determine which phrase in said first list of word phrase frequencies most likely constitutes said at least one company name.

7. The method of claim 1 further comprising:

automatically extracting company profile information from a web site associated with said at least one URL; and

automatically indexing and storing said extracted company profile information in said database such that it is relationally associated with said at least one URL.

8. The method of claim 7 wherein said company profile information comprises information pertaining to one or more of the following:

products;

services;

management team;

location;

size; and

age.

9. The method of claim 7 further comprising:

downloading content of a web site associated with said at least one URL address;

indexing and storing said content in said database such that is relationally associated with said at least one URL address; and

automatically and periodically updating at least a portion of said content with new content obtained from said web site associated with said at least one URL address.

10. The method of claim 9 wherein said step of automatically and periodically updating comprises calculating a change measure value based on differences between said content stored in said database and said new content, wherein if said change measure value exceeds a predetermined threshold value, said new content is stored so as to replace said at least a portion of said content in said database.

11. The method of claim 1 further comprising:

identifying at least one web page from a web site associated with said at least one URL address, wherein the at least one web page contains news information about a company associated with said web site;

extracting a URL address for said at least one web page;

indexing and storing said news information and said web page URL address such that they are relationally associated with said at least one URL address in said database; and

automatically and periodically updating said news information by accessing said web page using said web page URL address and determining whether new content is available.

12. The method of claim 11 wherein said step of determining whether new content is available comprises calculating a change measure value based on differences between said news information stored in said database and updated news information in said web page, wherein if said change measure value exceeds a predetermined threshold value, said updated news information is stored so as to replace said news information previously stored in said database.

13. A method of creating an industry database, comprising:

identifying a plurality of web sites meeting at least one search criteria;

automatically extracting URL addresses for each of said plurality of web sites;

automatically categorizing each of said plurality of web sites and their corresponding URL addresses in accordance with a predefined category structure comprising a plurality of categories; and

automatically indexing and storing each of said URL addresses in accordance with said predefined category structure in said database.

14. The method of claim 13 wherein said step of automatically categorizing comprises:

automatically downloading content from each of said plurality of web sites; and

automatically comparing said content from each of said web sites to at least one statistical model representative of at least one category in said predefined category structure.

15. The method of claim 14 further comprising calculating a confidence score based on said step of automatically comparing said content to said at least one statistical model.

16. The method of claim 14 wherein said statistical model represents word count characteristics of web site content previously categorized as belonging to said at least one category.

17. The method of claim 13 further comprising:

automatically extracting a plurality of company names each associated with a respective one of said URL addresses; and

automatically indexing and storing said plurality of company names under said predefined category structure in said database.

18. The method of claim 17 wherein said step of automatically extracting said plurality of company names comprises:

identifying and counting word phrase frequencies from content in said plurality of web sites, thereby creating a first list of word phrase frequencies;

for each of said web sites, identifying and counting word phrase frequencies found in each web site, thereby creating a second list of word phrase frequencies; and

for each of said web sites, comparing said first list of word phrase frequencies with said second list of word phrase frequencies to determine which phrase in said second list of word phrase frequencies most likely constitutes a respective company name.

19. The method of claim 13 further comprising:

automatically extracting company profile information from said plurality of web sites; and

automatically indexing and storing said extracted company profile information in said database such that it is relationally associated with respective ones of said plurality of web sites.

20. The method of claim 19 wherein said company profile information comprises information pertaining to one or more of the following:

products;

services;

management team;

location;

size; and

age.

21. The method of claim 19 further comprising:

downloading content from said plurality of web sites;

indexing and storing said content in said database such that is relationally associated with respective ones of said plurality of web sites; and

automatically and periodically updating at least a portion of said content with new content obtained from respective ones of said plurality of web site.

22. The method of claim 21 wherein said step of automatically and periodically updating comprises, for each respective web site, calculating a change measure value based on differences between said portion of said content previously stored in said database and new content found in said respective web site, wherein if said change measure value exceeds a predetermined threshold value, said new content is stored so as to replace said portion of said content previously stored in said database.

23. The method of claim 13 further comprising:

identifying at least one web page for each of said plurality of web sites, wherein the at least one web page contains news information about a respective company associated with each of said plurality of web sites;

extracting a URL address for each of said at least one web pages;

for each of said plurality of web sites, indexing and storing said respective news information and said respective web page URL addresses such that they are relationally associated with a respective one said plurality of web sites; and

for each of said plurality of web sites, automatically and periodically updating said respective news information by accessing said respective at least one web page and determining whether new content is available.

24. The method of claim 23 wherein said step of determining whether new content is available comprises calculating a change measure value based on differences between said respective news information stored in said database and updated news information in said respective at least one web page, wherein if said change measure value exceeds a predetermined threshold value, said updated news information is stored so as to replace said respective news information previously stored in said database.

25. An industry database, created in accordance with a process comprising the steps of:

conducting an Internet search for information meeting at least one search criteria;

creating a first list of URL addresses corresponding to web pages identified as a result of said Internet search;

unstemming said URL addresses in said first list to create a second list of URL addresses corresponding to unique web sites;

comparing said second list of URL addresses to URL addresses previously stored in said database;

deleting URL addresses from said second list that are duplicative of URL addresses previously stored in said database so as to create a third list of URL addresses;

automatically categorizing at least one URL address from said third list as belonging to a predefined category; and

automatically indexing and storing said at least one URL under said predefined category in said database.

26. The database of claim 25 wherein said step of automatically categorizing comprises:

selecting a subset of URL addresses from said third list so as to specify a training set for creating a statistical model;

downloading content from web sites corresponding to said subset of URL addresses;

creating a first word count list for each web site corresponding to said subset of URL addresses;

manually discarding at least one word determined to be a non-discriminating word from each of said first word count lists, creating a second word count list for each of said web sites;

manually classifying each URL address from said subset as either belonging to said predefined category or not belonging to said predefined category based on said content from corresponding web sites;

creating a statistical model representative of word count characteristics exhibited by web sites belonging to said predefined category and those web sites not belonging to said predefined category, based on said second word count lists;

validating said statistical model on said training set of web sites;

automatically downloading content from a web site corresponding to said at least one URL address from said third list; and

automatically comparing said content from said web site corresponding to said at least one URL address from said third list to said statistical model so as to automatically categorize said at least one URL as either belonging to or not belonging to said predefined category.

27. The database of claim 26 wherein said process further comprises calculating a confidence score based on said step of automatically comparing said content to said statistical model, wherein if said confidence score is below a threshold value, said at least one URL is presented to a human administrator for review.

28. The database of claim 26 wherein said statistical model further represents site structure characteristics of said web sites corresponding to said subset of URL addresses.

29. The database of claim 25 wherein said process further comprises automatically extracting at least one company name associated with said at least one URL and, thereafter, automatically indexing and storing said at least one company name under said predefined category in said database.

30. The database of claim 29 wherein said step of automatically extracting said at least one company name comprises:

identifying and counting word phrase frequencies from web site content associated with said at least one URL, thereby creating a first list of word phrase frequencies;

identifying and counting word phrase frequencies in content from a plurality of web sites associated with URL addresses in said second or third lists of URL addresses, thereby creating a second list of word phrase frequencies; and

comparing said first list of word phrase frequencies with said second list of word phrase frequencies to determine which phrase in said first list of word phrase frequencies most likely constitutes said at least one company name.

31. The database of claim 25 wherein said process further comprises:

automatically extracting company profile information from a web site associated with said at least one URL; and

automatically indexing and storing said extracted company profile information in said database such that it is relationally associated with said at least one URL.

32. The database of claim 31 wherein said company profile information comprises information pertaining to one or more of the following:

products;

services;

management team;

location;

size; and

age.

33. The database of claim 31 wherein said process further comprises:

downloading content of a web site associated with said at least one URL address;

indexing and storing said content in said database such that it is relationally associated with said at least one URL address; and

automatically and periodically updating at least a portion of said content with new content obtained from said web site associated with said at least one URL address.

34. The database of claim 33 wherein said step of automatically and periodically updating comprises calculating a change measure value based on differences between said portion of said content stored in said database and said new content, wherein if said change measure value exceeds a predetermined threshold value, said new content is stored so as to replace said at least a portion of said content in said database.

35. The database of claim 25 wherein said process further comprises:

identifying at least one web page from a web site associated with said at least one URL address, wherein the at least one web page contains news information about a company associated with said web site;

extracting a URL address for said at least one web page;

indexing and storing said news information and said web page URL address such that they are relationally associated with said at least one URL address in said database; and

automatically and periodically updating said news information by accessing said web page using said web page URL address and determining whether new content is available.

36. The database of claim 35 wherein said step of determining whether new content is available comprises calculating a change measure value based on differences between said news information stored in said database and updated news information in said web page, wherein if said change measure value exceeds a predetermined threshold value, said updated news information is stored so as to replace said news information previously stored in said database.

37. An industry database created in accordance with a process comprising the steps of:

identifying a plurality of web sites meeting at least one search criteria;

automatically extracting URL addresses for each of said plurality of web sites;

automatically categorizing each of said plurality of web sites and their corresponding URL addresses in accordance with a predefined category structure comprising a plurality of categories; and

automatically indexing and storing each of said URL addresses in accordance with said predefined category structure in said database.

38. The database of claim 37 wherein said step of automatically categorizing comprises:

automatically downloading content from each of said plurality of web sites; and

automatically comparing said content from each of said web sites to at least one statistical model representative of at least one category in said predefined category structure.

39. The database of claim 38 wherein said process further comprises calculating a confidence score based on said step of automatically comparing said content to said at least one statistical model.

40. The database of claim 38 wherein said statistical model represents word count characteristics of web site content previously categorized as belonging to said at least one category.

41. The database of claim 37 wherein said process further comprises:

automatically extracting a plurality of company names each associated with a respective one of said URL addresses; and

automatically indexing and storing said plurality of company names under said predefined category structure in said database.

42. The database of claim 41 wherein said step of automatically extracting said plurality of company names comprises:

identifying and counting word phrase frequencies from content in said plurality of web sites, thereby creating a first list of word phrase frequencies;

for each of said web sites, identifying and counting word phrase frequencies from web site content associated with said respective URL address, thereby creating a second list of word phrase frequencies; and

for each of said web sites, comparing said first list of word phrase frequencies with said second list of word phrase frequencies to determine which phrase in said second list of word phrase frequencies most likely constitutes a respective company name.

43. The database of claim 37 wherein said process further comprises:

automatically extracting company profile information from said plurality of web sites; and

automatically indexing and storing said extracted company profile information in said database such that it is relationally associated with respective ones of said plurality of web sites.

44. The database of claim 43 wherein said company profile information comprises information pertaining to one or more of the following:

products;

services;

management team;

location;

size; and

age.

45. The database of claim 43 wherein said process further comprises:

downloading content from said plurality of web sites;

indexing and storing said content in said database such that is relationally associated with respective ones of said plurality of web sites; and

automatically and periodically updating at least a portion of said content with new content obtained from respective ones of said plurality of web site.

46. The database of claim 45 wherein said step of automatically and periodically updating comprises, for each respective web site, calculating a change measure value based on differences between associated content previously stored in said database and new content found on said respective web site, wherein if said change measure value exceeds a predetermined threshold value, said new content is stored so as to replace said at least a portion of said associated content previously stored in said database.

47. The database of claim 37 wherein said process further comprises:

identifying at least one web page within said plurality of web sites, wherein the at least one web page contains news information about a respective company associated with a respective web site;

extracting a URL address for said at least one web page;

indexing and storing said respective news information and said respective web page URL address such they are relationally associated with a respective one said plurality of web sites; and

automatically and periodically updating said respective news information by accessing said respective at least one web page and determining whether new content is available.

48. The database of claim 47 wherein said step of determining whether new content is available comprises calculating a change measure value based on differences between said respective news information stored in said database and updated news information in said respective at least one web page, wherein if said change measure value exceeds a predetermined threshold value, said updated news information is stored so as to replace said respective news information previously stored in said database.

49. A database system comprising:

a relational database containing a plurality of URL addresses for a plurality web sites indexed and stored in accordance with a predefined category structure; and

a company directory search engine for automatically retrieving new URL addresses for new web sites, automatically categorizing said new URL addresses and new web sites, and storing at least a subset of said new URL addresses in said relational database in accordance with said predefined category structure.

50. The database system of claim 49 further comprising a BioField search engine for automatically downloading content from said plurality of web sites, automatically categorizing said content and storing said content in said relational database in accordance with said predefined category structure.

51. The database system of claim 50 wherein said BioField search engine also automatically and periodically updates at least a portion of said content with new content obtained from at least one of said plurality of web sites.

52. The database system of claim 49 further comprising a BioNews search engine that automatically identifies web pages within said plurality of web sites and indexes and stores URL address for said web pages in said database, wherein said web pages contain news pertaining to respective companies associated with respective web sites, wherein the BioNews search engine automatically downloads news content from said identified web pages, stores said news content in said database in accordance with said predefined category structure, and periodically and automatically updates said news content with new information obtained from one or more of said identified web pages.

53. The database system of claim 49 further comprising an Opportunity search engine that automatically and periodically searches preselected web pages having URL addresses stored and indexed in said database in accordance with said predefined category structure, wherein said preselected web pages contain information pertaining to opportunities for companies belonging to an industry, and wherein said Opportunity search engine automatically downloads, categorizes, indexes and stores content from said web pages and periodically updates this content with new content obtained from said web pages.

54. The database system of claim 53 further comprising a technology alert module for receiving a plurality of user queries relating to business opportunities and periodically comparing said user queries with one another as well as opportunity information stored and indexed in said relational database to determine if there is a potential match between two or more user queries or between a user query and one or more entries of opportunity information stored and indexed in the database, wherein said technology alert module sends a message to appropriate users if a potential match is found.

55. The database system of claim 49 further comprising a job module for automatically and periodically identifying and extracting job opening information from said plurality of web sites, indexing and storing said information in said relational database, and comparing said information with requests received from users of said system to determine if there is a potential match between one of said requests and said job opening information from one or more of said plurality of web sites.

56. The database system of claim 49 further comprising a start-up module for receiving a plurality of proposals from member companies, wherein the start-up module automatically categorizes and indexes each of said plurality of proposal in accordance with said predefined category structure, thereby allowing focused searches to be performed by other member companies desiring to view only a subset of said plurality of proposals indexed under one or more desired categories in said predefined category structure.

57. The database system of claim 49 wherein:

said relational database further contains company profile information extracted from said plurality of web sites, wherein said company profile information is indexed and stored in said relational database in accordance with said predefined category structure;

wherein at least a subset of the entries for said company profile information stored in the relational database are “

linked”

to one another such that changes to one entry trigger changes to one or more other linked entries, in accordance with a specified linking logic; and

wherein if one of said company profile entries are updated with new information, said one or more other linked entries are automatically updated in accordance with said specified linking logic.

58. The database system of claim 57 wherein said linked company profile information includes the following information types:

management team, contact information, new financing, M&

A transactions, and new partners.

59. A method of providing information responsive to user queries, comprising:

storing in a database, information extracted from a plurality of web sites, wherein said information is automatically categorized and indexed in accordance with a predefined category structure and wherein said information includes a plurality of URL addresses corresponding to said plurality of web sites;

receiving a user query;

executing a search engine in response to said user query wherein said search engine searches a subset of said stored information extracted from a subset of said plurality of web sites, wherein said subset of information is selected based on corresponding category indices that match said use query; and

searching said subset of web sites to find additional information responsive to said user query.

60. A database system for providing information responsive to user queries, comprising:

a database for storing information extracted from a plurality of web sites, wherein said information is automatically categorized and indexed in accordance with a predefined category structure and wherein said information includes a plurality of URL addresses corresponding to said plurality of web sites;

a user interface module for receiving a user query; and

a server computer for executing said user interface module and a search engine in response to said user query, wherein said search engine searches a subset of a said stored information extracted from a subset of said plurality of web sites, wherein said subset of information is selected based on corresponding category indices matching said use query, and wherein said search engine subsequently searches said subset of web sites to find additional information responsive to said user query.