Probabilistic method for natural language processing and for encoding free-text data into a medical database by utilizing a Bayesian network to perform spell checking of words

• US 6,292,771 B1
• Filed: 09/30/1998
• Issued: 09/18/2001
• Est. Priority Date: 09/30/1997
• Status: Expired due to Term

• Current Assignee: Intermountain Intellectual Asset Management LLC
• Original Assignee: IHC Health Services Incorporated (IHC Hospitals, Inc.)
• Inventors: Gundersen, Michael L., Koehler, Spencer B., Haug, Peter J., Van Bree, Rudy E., Christensen, Lee M.
• Primary Examiner(s): Thomas, Joseph
• Application Number:
  US09/164,048
• Time in Patent Office: 1,084 Days
• Field of Search: 704/9, 704/1, 704/10, 704/251, 704/255, 704/256, 704/257, 705/2, 705/3, 705/4
• US Class Current: 704/9
• CPC Class Codes: G06F 40/10 Text processing natural language analysis G06F40/20; semantic analysis G06F40/30; processing or translation of natural language G06F40/40
• CPC Class Codes: G06F 40/211 Syntactic parsing, e.g. based on context-free grammar [CFG] or unification grammars
• CPC Class Codes: G06F 40/216 using statistical methods
• CPC Class Codes: G06F 40/232 Orthographic correction, e.g. spell checking or vowelisation
• CPC Class Codes: G06F 40/30 Semantic analysis

Abstract

A natural language understanding system is described which provides for the generation of concept codes from free-text medical data. A probabilistic model of lexical semantics, in the preferred embodiment of the invention implemented by means of a Bayesian network, is used to determine the most probable concept or meaning associated with a sentence or phrase. The inventive method and system includes the steps of checking for synonyms, checking spelling, performing syntactic parsing, transforming text to its “deep” or semantic form, and performing a semantic analysis based on a probabilistic model of lexical semantics. In the preferred embodiment of the invention, spell checking and transformational processing as well as semantic analysis make use of semantic probabilistic determinations.

Claims

1. A method for encoding free-text data, comprising:

(A) receiving free-text data, wherein said free text data, wherein said free-text data includes;

words, a grammar, a syntax and a semantic relationship between said words;

(B) checking for synonyms of said words within said received free-text data;

(C) checking spelling of said words within said received free-text data wherein said checking spelling further comprises;

(1) identifying a target word within said free-text data for spell checking;

(2) identifying a set of words known to a Bayesian network, wherein said identified set of words are those which said target word could be a misspelling, wherein said identification is based on transformations of said target word from said free-text data to create a list of candidate spellings of said target word;

(3) binding said set of candidate spellings to said words known to said Bayesian network;

(4) identifying a best candidate spelling from said list of candidate spelling based on probability values produced by said Bayesian network; and

(5) replacing said target word with said best candidate spelling;

(D) parsing said syntax of said received free-text data;

(E) transforming said grammar of said received free-text data;

(F) analyzing said semantic relationship within said received free-text data on the basis of a probabilistic model of lexical semantics which relates said words to one or more concepts;

(G) creating an encoded representation of said received free-text data; and

(H) writing said encoded representation into a medical database.

2. A method for encoding free-text data, comprising:

(A) receiving free-text data, wherein said free text data, wherein said free-text data includes;

words, a grammar, a syntax and a semantic relationship between said words;

(B) checking for synonyms of said words within said received free-text data;

(C) checking spelling of said words within said received free-text data, wherein said checking spelling further comprises;

(1) identifying a target word within said free-text data for spell checking;

(2) identifying a set of words known to a Bayesian network, wherein said identified set of words are those which said target word could be a misspelling, wherein said identification is based on transformations of said target word from said free-text data to create a list of candidate spellings of said target word;

(3) binding said set of candidate spellings to said words known to said Bayesian network;

(4) identifying a best candidate spelling from said list of candidate spelling based on probability values produced by said Bayesian network, wherein said identifying a best candidate spelling employs a probabilistic analysis of said candidate spellings; and

(5) replacing said target word with said best candidate spelling;

(D) parsing said syntax of said received free-text data;

(E) transforming said grammar of said received free-text data;

(F) analyzing said semantic relationship within said received free-text data on the basis of a probabilistic model of lexical semantics which relates said words to one or more concepts;

(G) creating an encoded representation of said received free-text data; and

(H) writing said encoded representation into a medical database.