Method for extracting knowledge from online documentation and creating a glossary, index, help database or the like

• US 5,799,268 A
• Filed: 09/28/1994
• Issued: 08/25/1998
• Est. Priority Date: 09/28/1994
• Status: Expired due to Term

• Current Assignee: Apple Inc.
• Original Assignee: Apple Computer Incorporated (Apple Inc.)
• Inventors: Boguraev, Branimir K.
• Primary Examiner(s): Weinhardt, Robert A.
• Assistant Examiner(s): Hughet, William N.
• Application Number:
  US08/314,589
• Time in Patent Office: 1,427 Days
• Field of Search: 395/759, 395/760, 395/761, 704/9, 704/10, 707/900
• US Class Current: 704/9
• CPC Class Codes: G06F 40/289 Phrasal analysis, e.g. finite state techniques or chunking
• CPC Class Codes: G06F 40/30 Semantic analysis
• CPC Class Codes: Y10S 707/99943 Generating database or data structure, e.g. via user interface

Abstract

A method involving computer-mediated linguistic analysis of online technical documentation to extract and catalog from the documentation knowledge essential to, for example, creating a online help database useful in providing online assistance to users in performing a task. The method comprises stripping markup tags from the documentation, linguistically analyzing and annotating the text, including the steps of morphologically and lexically analyzing the text, disambiguating between possible parts-of-speech for each word, and syntactically analyzing and labeling each word. The method further comprises the steps of combining the linguistically analyzed, annotated, and labeled text and previously stripped markup information into a merged file, mining the merged file for domain knowledge, including the steps of identifying and creating a list of technical terminology, mining the merged file for manifestations of domain primitives and maintaining a list of manifestations of such domain primitives in an observations file, analyzing the discourse context of each sentence or phrase in the merged file, analyzing the frequency of manifestations of domain primitives in the observations file to determine those that are important, expanding the list of key terms by searching for terms sanctioned by a domain primitive deemed important in the previous step, and searching the merged file for larger relations by searching for particular lexico-syntactic patterns involving key terms and manifestations of domain primitives previously identified. The method further comprises the steps of structuring the knowledge thus mined and building a domain catalog.

Claims

1. In a computer system having access to online documentation, a method of extracting knowledge from said online documentation, comprising the steps of:

a) linguistically analyzing and annotating text of said online documentation to create a linguistically analyzed and annotated text;

b) mining said linguistically analyzed and annotated text for text representative of said online documentation, including the steps of;

i) searching for syntactic patterns indicative of key terms and maintaining a list of said key terms;

ii) searching for syntactic patterns indicative of manifestations of a domain primitive involving one of said key terms and maintaining a list of said manifestations; and

iii) analyzing said list of said manifestations to determine said manifestations that are representative of said online documentation on the basis of frequency of their occurrence; and

c) combining said list of said key terms and said list of said manifestations that are representative of said online documentation in a domain catalog.

2. The method of claim 1, wherein said online documentation comprises online technical documentation.

3. The method of claim 1, wherein said key terms comprise technical terms and said list of said key terms comprises a list of said technical terms.

4. In a computer system having access to online technical documentation, a method of extracting knowledge from said online technical documentation, comprising the steps of:

a) linguistically analyzing and annotating text of said online technical documentation to create a linguistically analyzed and annotated text, including the steps of;

i) lexically and morphologically analyzing said text;

ii) disambiguating between possible parts-of-speech for each word of said text and identifying a syntactic function for each word of said text;

b) mining said linguistically analyzed and annotated text for text representative of said online technical documentation, including the steps of;

i) searching for syntactic patterns indicative of technical terms and maintaining a list of said technical terms and the frequency with which said technical terms occur in said text;

ii) searching for syntactic patterns indicative of manifestations of a domain primitive involving one of said technical terms and maintaining a list of said manifestations;

iii) analyzing said list of said manifestations to determine said manifestations that are representative of said online technical documentation on the basis of frequency of their occurrence;

iv) expanding said list of technical terms by searching said list of said manifestations for syntactic patterns involving additional terms not presently in said list of said technical terms that are adjunct to said manifestations and,

v) searching said online technical documentation for lexico-syntactic patterns indicative of larger relations involving said technical terms, said additional terms and said manifestations; and

c) combining said technical terms, said additional terms, said manifestations, and said larger relations in a domain catalog.

5. The method of claim 4, wherein said online technical documentation is stored in an American Standard Code for Information Interchange, hereinafter referred to as ASCII, data file conforming with a standard internal representation for document structure.

6. In a computer system having access to an online technical document, a method of extracting knowledge from said online technical document, comprising the steps of:

a) translating an American Standard Code for Information Interchange, hereinafter referred to as ASCII, data file having stored therein said online technical document and information regarding a proprietary internal representation for said online technical document to a second ASCII data file having stored therein said online technical document and information regarding a standard internal representation for said online technical document;

b) separating text of said online technical document from said information regarding said standard internal representation for said online document;

c) linguistically analyzing and annotating said text of said online technical documenta to create a linguistically analyzed and annotated text, including the steps of;

i) lexically and morphologically analyzing said text to determine possible lexical and morphological features;

ii) disambiguating between possible parts-of-speech for each word of said text and identifying a syntactic function for each word of said text;

d) labeling each word of said text and said annotations;

e) combining said linguistically analyzed and annotated text with said information regarding said standard internal representation for said online document into a merged file;

f) mining said merged file for text representative of said online technical document, including the steps of;

i) identifying key terms, including the steps of;

1) searching for syntactic patterns indicative of technical terms; and

2) maintaining a list of said key terms comprising said technical terms and the frequency with which said technical terms occur;

ii) searching for syntactic patterns indicative of manifestations of a domain primitive involving one of said key terms and maintaining a list of said manifestations and the frequency with which said manifestations occur;

iii) discourse analyzing each phrase in said merged file to resolve conjunctions and antecedent basis of pronouns to more accurately record said manifestations and the frequency with which said manifestations occur;

iv) analyzing said list of said manifestations to determine said manifestations that are representative of said online document on the basis of said frequency with which said manifestations occur;

v) expanding said list of key terms by searching said list of said manifestations for syntactic patterns involving additional terms not presently in said list of key terms that are adjunct to one of said manifestations; and

vi) searching said online technical document for lexico-syntactic patterns indicative of larger relations involving one of said key terms and one of said manifestations and maintaining a list of said larger relations; and

g) combining said key terms, said manifestations, and said larger relations in a domain catalog.

7. The method of claim 6, wherein said frequency with which said technical terms occur in said merged file is weighted according to said information regarding said standard internal representation for said online document.

8. The method of claim 6, wherein analyzing said list of manifestations to determine said manifestations that are representative of said online document on the basis of said frequency with which said manifestations occur is weighted according to said information regarding said standard internal representation for said online document.

9. The method of claim 6 wherein said domain primitive comprises terms, relations, or properties syntactically related to said key terms.

10. The method of claim 6 wherein combining said key terms, said manifestations, and said larger relations in a domain catalog comprises the steps of:

a) clustering said key terms for similarity of use on the basis of repeated manifestations of said manifestations of a domain primitive occurring in a similar syntactic context;

b) clustering said key terms on the basis of proximity of said key terms in said merged file with respect to other said key terms; and

c) indexing said domain catalog by said key terms such that each of said key terms is an index therein, and incorporating said manifestations and said larger relations according to said indexed key terms.

11. In a computer system having access to source documentation, a method of designing the content of a help database, comprising the steps of:

a) linguistically analyzing and annotating text of said source documentation to create a linguistically analyzed and annotated text;

b) mining said linguistically analyzed and annotated text for text representative of said source documentation, including the steps of;

i) searching for syntactic patterns indicative of key terms and maintaining a list of said key terms;

ii) searching for syntactic patterns indicative of manifestations of a primitive involving one of said key terms and maintaining a list of said manifestations; and

iii) analyzing said list of said manifestations to determine said manifestations that are representative of said source documentation;

c) combining said list of said key terms and said list of said manifestations that are representative of said source documentation in a catalog; and

d) using said catalog to create said content of said help database.

12. The method of claim 11, wherein said content comprises a file containing said list of said key terms and said list of said manifestations that are representative of said source documentation.

13. In a desktop publishing system, having access to a reference manual, a method of designing an index for said reference manual, comprising the steps of:

a) linguistically analyzing and annotating text of said reference manual to create a linguistically analyzed and annotated text;

b) mining said linguistically analyzed and annotated text for text representative of said reference manual, including the steps of;

i) searching for syntactic patterns indicative of index terms and maintaining a list of said index terms;

ii) searching for syntactic patterns indicative of manifestations of a primitive involving one of said index terms and maintaining a list of said manifestations; and

iii) analyzing said list of manifestations to determine said manifestations that are representative of said reference manual on the basis of the frequency of their occurrence;

c) combining said list of said index terms and said list of said manifestations that are representative of said reference manual in an index file; and

d) using said index file to create said index for said reference manual.

14. In a desktop publishing system, having access to a reference manual, a method of designing a glossary for said reference manual, comprising the steps of:

a) linguistically analyzing and annotating text of said reference manual to create a linguistically analyzed and annotated text;

b) mining said linguistically analyzed and annotated text for text representative of said reference manual, including the steps of;

i) searching for syntactic patterns indicative of glossary terms and maintaining a list of said glossary terms;

ii) searching for syntactic patterns indicative of manifestations of a primitive involving one of said glossary terms and maintaining a list of said manifestations; and

iii) analyzing said list of said manifestations to determine said manifestations that are representative of said reference manual on the basis of their frequency of occurrence;

c) combining said list of said glossary terms and said list of said manifestations that are representative of said reference manual in a glossary file; and

d) using said glossary file to create said glossary for said reference manual.

15. In a computer system having access to a source code document, a method for designing a library of variables and procedure calls for said source code document, comprising the steps:

a) linguistically analyzing and annotating a source code of said source code document to create a linguistically analyzed and annotated source code;

b) mining said linguistically analyzed and annotated source code for source code representative of said variables and procedure calls, including the steps of;

i) searching for syntactic patterns indicative of key terms comprising variables and procedure calls, and maintaining a list of said key terms;

ii) searching for syntactic patterns indicative of manifestations of a primitive involving one of said key terms and maintaining a list of said manifestations; and

iii) analyzing said list of said manifestations to determine said manifestations that are representative of said variables and procedure calls on the basis of frequency of their occurrence;

c) combining said list of said key terms and said list of said manifestations that are representative of said variables and procedure calls in a library file; and

d) using said library file to create said library of variables and procedure calls.