Conceptual world representation natural language understanding system and method

• US 8,442,814 B2
• Filed: 03/01/2011
• Issued: 05/14/2013
• Est. Priority Date: 07/12/2002
• Status: Expired due to Term

• Current Assignee: Nuance Communications, Inc. (Microsoft Corporation)
• Original Assignee: Nuance Communications, Inc. (Microsoft Corporation)
• Inventors: Ceusters, Werner, O'Donnell, Mick, Montyne, Frank, Coppens, Frederik, Van Mol, Maarten
• Primary Examiner(s): He, Jialong
• Application Number:
  US13/037,478
• Publication Number:
  US 20110179032A1
• Time in Patent Office: 805 Days
• Field of Search: 704/6, 704/7, 704/9, 704/4, 707/737, 707/749, 706/46
• US Class Current: 704/9
• CPC Class Codes: G06F 16/35 Clustering; Classification
• 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: G06F 40/40 Processing or translation of natural language natural language analysis G06F40/20; semantic analysis G06F40/30
• CPC Class Codes: G06N 5/02 Knowledge representation; Symbolic representation

Abstract

A Natural Language Understanding system is provided for indexing of free text documents. The system according to the invention utilizes typographical and functional segmentation of text to identify those portions of free text that carry meaning. The system then uses words and multi-word terms and phrases identified in the free to text to identify concepts in the free text. The system uses a lexicon of terms linked to a formal ontology that is independent of a specific language to extract concepts from the free text based on the words and multi-word terms in the free text. The formal ontology contains both language independent domain knowledge concepts and language dependent linguistic concepts that govern the relationships between concepts and contain the rules about how language works. The system according to the current invention may preferably be used to index medical documents and assign codes from independent coding systems, such as, SNOMED, ICD-9 and ICD-10. The system according to the current invention may also preferably make use of syntactic parsing to improve the efficiency of the method.

Claims

1. A method of processing free text documents for indexing, said method comprising:

typographically segmenting a free text document using a computer-based ontology management system, said typographically segmenting comprising;

delimiting said free text document into words, sentences, titles, list items and paragraphs based on character patterns in said free text document;

functionally segmenting said free text document using the computer-based ontology management system, said functionally segmenting comprising;

grouping words into multi-word terms, segmenting said sentences into clause-phrase segments, and grouping words into noun phrases, wherein said grouping words into multi-word terms is accomplished by identifying at least two adjacent words;

re-writing at least one of said at least two adjacent words to generate a pairing of at least two adjacent words containing at least one re-written word;

searching a lexicon of terms for said pairing of at least two adjacent words containing at least one re-written word;

if said pairing of at least two adjacent words containing at least one re-written word is found in said lexicon, replacing said pairing of at least two adjacent words with said pairing of at least two adjacent words containing at least one re-written word; and

tagging said pairing of at least two adjacent words containing at least one re-written word as a multi-word term.

2. The method according to claim 1, wherein

said delimiting of said free text document into words is accomplished by recognizing spaces and punctuation marks between characters in said free text,

said delimiting of said free text document into sentences is accomplished by recognizing punctuations following a series of words, wherein said punctuations are defined as ending sentences, and

said delimiting of said free text document into paragraphs is accomplished by recognizing a paragraph marker in said free text document.

3. The method according to claim 1, wherein said grouping words into multi-word terms is further accomplished by:

pairing said at least two adjacent words,

searching said lexicon of terms for said pairing of at least two adjacent words, and

if said pairing is found on said lexicon of terms, tagging said pairing as a multi-word term.

4. The method according to claim 1, wherein said segmenting said sentences into clause-phrase segments comprises:

identifying a first segment and a second segment in a sentence, wherein said first segment and said second segment are split by a marker that signals the start of a new clause or phrase, and

tagging said first segment as a first clause or phrase and tagging said second segment as a second clause or phrase.

5. The method according to claim 4, wherein said marker that signals the start of a new clause or phrase is selected from the group consisting of:

and, but, or, “

,”

, “

;

”

, although, however, therefore, because, since, during, until, which, if, except, who, while, when, where, with, without, “

to avoid”

, and “

to the point”

,

with the following proviso;

if said first segment and said second segment are split by “

and”

or “

or” and

said first segment ends in a noun phrase and said second segment begins in a noun phrase, said first segment and said second segment are tagged as a single clause or phrase;

if said first segment and said second segment are split by “

,” and

said first segment ends in with nominal word and said second segment begins in with nominal word, said first segment and said second segment are tagged as a single clause or phrase;

if said first segment and said second segment are split by “

,” and

said first segment is an adverb, said first segment and said second segment are tagged as a single clause or phrase; and

if said second segment comprises “

etc.”

, said first segment and said second segment are tagged as a single clause or phrase.

6. The method according to claim 1, further comprising:

identifying negating words in said free text.

7. The method according to claim 6, wherein said negating words are selected from the group consisting of:

not, no, without, zero, non, nor, avoid, absence, denies, deny, denied, never, won'"'"'t, shouldn'"'"'t, wouldn'"'"'t, couldn'"'"'t, can'"'"'t, “

with no” and

“

ruled out”

.

8. The method according to claim 7, wherein clauses or phrases containing negating words are tagged as negating text and ignored in further processing.

9. The method according to claim 1, further comprising:

identifying modalizing words in said free text.

10. The method according to claim 9, wherein said modalizing words are selected from the group consisting of:

might, may, would, could, should, possibly, probably, can presumed, prefers, prefer, preferred, preferably, wants, wanted, wanting, desires, desired, desire, desiring, likely, unlikely, encourage, encouraged, if, maybe, questionable and suggestive.

11. The method according to claim 9, wherein clauses or phrases containing modalizing words are tagged as modalised text.

12. The method according to claim 11, further comprising identifying modalizing words adjacent to negating words, wherein clauses or phrases containing modalizing words adjacent to negating words are tagged as modalised text.

13. The method according to claim 1, further comprising:

grouping said paragraphs into functional sections.

14. The method according to claim 13, further comprising:

labeling said functional sections by topic.

15. The method according to claim 1, further comprising:

syntactically parsing said free text document.

16. The method according to claim 15, wherein said syntactic parsing is performed using dependency grammar.