Apparatus and method for retrieving and grouping images representing text files based on the relevance of key words extracted from a selected file to the text files

• US 5,598,557 A
• Filed: 09/22/1992
• Issued: 01/28/1997
• Est. Priority Date: 09/22/1992
• Status: Expired due to Term

• Current Assignee: Nuance Communications, Inc. (Microsoft Corporation)
• Original Assignee: Caere Corp. (Mexico) (Microsoft Corporation)
• Inventors: Barnes, Michael R., Miller, Lawrence G., Doner, Christopher G., Emmons, Ian D.
• Primary Examiner(s): Black, Thomas G.
• Assistant Examiner(s): Choules, Jack M.
• Application Number:
  US07/948,669
• Time in Patent Office: 1,589 Days
• Field of Search: 395/600, 395/159, 364/419.08, 364/419.19
• US Class Current: 1/1
• CPC Class Codes: G06F 16/3346 using probabilistic model
• CPC Class Codes: Y10S 707/99935 Query augmenting and refining, e.g. inexact access

Abstract

An apparatus for searching and retrieving files in a database without a user being required to provide keywords or query terms. A user first selects and opens a reference file. A natural language recognition algorithm is used to determine the subject words of the selected file. Next, a statistical comparison between the subject words and the contents of files in a database is performed. Based on the statistical comparison, files are assigned weighted relevancies. Relevant files are prioritized and displayed to the user in groups. The groups are formed based on the retrieved files relevance to specific subject works of the selected file. The groups of retrieved files are displayed in associating with the subject word they are relevant to.

Claims

1. A computer system for searching and retrieving files in a database comprising:

a bus;

an input device coupled to said bus for selecting a first reference file;

a storage area coupled to said bus having stored therein a program;

a processor coupled to said bus for executing said program, said program including

an algorithm for determining significant words of text comprising said first file, and

a comparison algorithm for comparing said significant words of said first reference file and text of a second file in said database; and

a display device, coupled to said bus, for displaying a plurality of said significant words from said first reference file and one or more images corresponding to retrieved files, wherein said one or more images are grouped based upon the retrieved files'"'"' relevance to said plurality of said significant words, wherein selecting one of said images designates the corresponding retrieved file as a second reference file for initiating a second search and retrieval of files in said database which are relevant to said second reference file.

2. The apparatus of claim 1, wherein said program is further comprised of:

an algorithm for repeatedly performing statistical comparisons between said significant words of said first reference file and text of other files of said database and assigning a weight to each of said other files;

a ranking algorithm for ranking said other files according to said weights and displaying a list of said ranked files.

3. The apparatus of claim 2, wherein said significant words include subject words.

4. The apparatus of claim 2, wherein said significant words include direct objects.

5. The apparatus of claim 2, wherein said significant words include bold, underlined, and italicized words.

6. The apparatus of claim 2, wherein said significant words are weighted before performing said comparison.

7. The apparatus of claim 2 is further comprised of a search algorithm for searching, retrieving, and ranking files in said database based on keywords.

8. The apparatus of claim 1, wherein said algorithm for determining said significant words is comprised of:

a parsing algorithm for parsing text of said reference file into sentences;

a filtering algorithm for disregarding inconsequential words in a sentence;

a natural language processing algorithm for determining which of said words in said sentence are significant based on parts of speech for said significant words and a set of heuristic grammar rules.

9. The apparatus of claim 8, wherein said significant words include subject words.

10. The apparatus of claim 8, wherein said significant words include direct objects.

11. The apparatus of claim 8, wherein said significant words include bold, underlined, and italicized words.

12. The apparatus of claim 8, wherein said significant words are weighted before performing said comparison.

13. The apparatus of claim 8 is further comprised of a search algorithm for searching, retrieving, and ranking files in said database based on keywords.

14. The apparatus of claim 1, wherein said comparison is a function of a frequency of significant words in said second file, a function of a distribution of said significant words within said database, and is normalized by a function of a length of said second file.

15. The apparatus of claim 14, wherein said significant words include subject words.

16. The apparatus of claim 14, wherein said significant words include direct objects.

17. The apparatus of claim 14, wherein said significant words include bold, underlined, and italicized words.

18. The apparatus of claim 14, wherein said significant words are weighted before performing said comparison.

19. The apparatus of claim 14 is further comprised of a search algorithm for searching, retrieving, and ranking files in said database based on keywords.

20. An apparatus for searching, retrieving, and displaying files in a database comprising:

a computer system for selecting and opening a reference file;

a search algorithm executed by said computer system for searching for and retrieving files from said database which are related to said reference file based on comparing significant words of said reference file and text of other files residing in said database;

an algorithm for categorizing and prioritizing said retrieved files into a plurality of groups, wherein each of said plurality of groups corresponds to a commonly used subject word in said reference file; and

a display for displaying said commonly used subject words, bit-mapped images representing said prioritized files corresponding to commonly used words, and a bit-mapped image representing said reference file, said bit-mapped images being scaled representations of document cover sheets in said files, wherein selecting a bit-mapped image results in designating its corresponding file as the reference file and initiates a search and retrieval based on said designated reference file.

21. The apparatus of claim 20 wherein said searching and retrieving means is comprised of:

a means for performing a statistical comparison between said significant words of said reference file and files residing in said database;

a means for retrieving files from said database based on said statistical comparison.

22. A method for searching and retrieving files in a database comprising the steps of:

(a) selecting a first reference file;

(b) determining significant words of text comprising said first reference file;

(c) comparing said significant words of said first reference file and text of a second file in said database;

(d) determining whether to retrieve said second file based on said step of comparing;

(e) retrieving said second file if said step of determining indicates said second file should be retrieved;

(f) displaying a plurality of said significant words;

(g) displaying one or more images corresponding to retrieved files, wherein said one or more images are grouped based upon the retrieved files'"'"' relevance to said plurality of said significant words; and

(h) upon selection of one of said images, designating the corresponding retrieved file as a second reference file and initiating a second search and retrieval of files in said database relevant to said second reference file.

23. The method of claim 22, further comprising the steps of:

repeatedly performing comparisons between said significant words of said first reference file and other files of said database;

assigning a weight to each of said other files based on said comparison;

ranking said other files according to said weights;

displaying a list of ranked files.

24. The method of claim 23, wherein said determining step is comprised of the steps of:

parsing text of said first reference file into sentences;

disregarding inconsequential words in said sentences;

determining which of said words in said sentences are significant based on parts of speech for said significant words and heuristic grammar rules.

25. (Amended) The method of claim 24, wherein said comparison is a statistical comparison as a function of frequency of significant words in said second file and a distribution of said significant words within said database and is normalized by a function of a length of said second file.

26. The method of claim 25, wherein said significant words include subject words.

27. The method of claim 25, wherein said significant words include direct objects.

28. The method of claim 25, wherein said significant words include bold, underlined, and italicized words.

29. The method of claim 25, further comprising the step of weighting said significant words before performing said statistical comparison.

30. The method of claim 25, further comprising the steps of searching, retrieving, and ranking files in said database according to keywords.