Automated word substitution using numerical rankings of structural disparity between misspelled words & candidate substitution words

• US 4,783,758 A
• Filed: 02/05/1985
• Issued: 11/08/1988
• Est. Priority Date: 02/05/1985
• Status: Expired due to Term

• Current Assignee: Vantage Technology Holdings LLC
• Original Assignee: Houghton Mifflin Harcourt Publishing Company (Houghton Mifflin Harcourt Co. (Private Equity))
• Inventors: Kucera, Henry
• Primary Examiner(s): Williams, Jr, Archie E.
• Application Number:
  US06/699,202
• Time in Patent Office: 1,372 Days
• Field of Search: 364/200 MS File, 364/900 MS File, 364/419, 400/61, 400/63, 400/83
• US Class Current: 715/210
• CPC Class Codes: G06F 40/232 Orthographic correction, e.g. spell checking or vowelisation

Abstract

A spelling correction system compares a correctly spelled word with an incorrectly spelled word to determine the degree of substitutability. If the system determines that the words are highly similar, the system flags the correct word as exclusively substitutable for the incorrect word. If the system determines the words are of moderate similarity, the correct word is flagged as a possible substitute for the incorrect word.

Claims

1. In apparatus for automated linguistic expression substitution, the improvement comprising

A. first input means for inputting, into said apparatus, a suspect expression signal representative of a linguistic expression consisting of characters,

B. second input means for inputting, into said apparatus, an alternate expression signal representative of a permissible linguistic expression consisting of characters,

C. comparison means, connected to said first and second input means, for comparing said suspect expression signal with said alternate expression signal and for producing a disparity signal numerically representative of differences between a spelling of the linguistic expression represented by said suspect expression signal and a spelling of the linguistic expression represented by said alternate expression signal, said comparison means including means for detecting disparity types representative of

(i) transposition of character order,

(ii) deletion of a character,

(iii) an unmatched character, and

(iv) differences in expression character length, occurring between said compared expression signals, said comparison means further including means responsive to the detection of said disparity type for producing a signal indicative of the numerically-weighted structural significance of that type, and

D. selection means, connected to said comparison means and responsive to a numerical value represented by said disparity signal, for selectively determining said alternate expression to be substitutable for said suspect expression and for producing an output signal indicative thereof.

2. In apparatus according to claim 1, the further improvement whereby said comparison means includes means for producing said disparity signal to be numerically representative of the type and magnitude of differences between said suspect expression signal and said alternate expression signal.

3. In apparatus according to claim 1, the further improvement whereby said comparison means includes means for detecting a disparity type representative of transposition of character order, with respect to said alternate expression signal, within said suspect expression signal.

4. In apparatus according to claim 1, the further improvement whereby said comparison means includes means for detecting a disparity type representative of the deletion of a character, with respect to said alternate expression signal, from said suspect expression signal.

5. In apparatus according to claim 1, the further improvement whereby said comparison means includes means for detecting a disparity type representative of an unmatched character, with respect to said alternate expression signal, within said suspect expression signal.

6. In apparatus according to claim 1, the further improvement whereby said comparison means includes means for detecting a disparity type resulting from differences in the number of characters represented by said suspect expression signal and by said alternate expression signal.

7. In apparatus according to claim 1, the further improvement whereby said comparison means includes means for detecting like characters represented in said suspect expression signal and said alternate expression signal.

8. In apparatus according to claim 1, the further improvement whereby said comparison means further includes means for computing a disparity value as a summation of values indicated by each said significance signal.

9. Apparatus for automated linguistic expression substitution having a keyboard for accepting expressions and converting them into signals representative thereof, a monitor for graphic display of signals representative of symbols and linguistic expressions, a processing unit for processing signals representative of lingiuistic expressions, and a storage unit for storing signals representative of linguistic expressions, and further having the improvement comprising

A. first input means for inputting from at least one of said keyboard, said processing unit, and said storge unit, a suspect expression signal representative of a linguistic expression,

B. second input means for inputting, from at least one of said processing unit or said storage unit, alternate expression signals representative of one or more permissible linguistic expressions,

C. comparison means, connected to said first and second input means, for comparing said suspect expression signal with each said alternate expression signal and for producing a disparity signal numerically representative of differences between a spelling of the linguistic expression represented by said suspect expression signal and a spelling of the linguistic expression represented by each said alternate expression signal, said comparison means including means for detecting disparity types representative of

(i) transposition of character order,

(ii) deletion of a character,

(iii) an unmatched character, and

(iv) differences in expression character length, occurring between said compared expression signals, and further includes means responsive to the detection of said disparity type for producing a signal indicative of the numerically-weighted structural significance of that type, and

D. selection means, connected to said comparison means and responsive to a numerical value represented by said disparity signals, for selecting the alternate expression signal, if any, to replace said suspect expression and for producing an output signal indicative thereof.

10. Apparatus according to claim 9, having the further improvement whereby said comparison means includes means for producing said disparity signal to be numerically representative of the type and magnitude of differences between said suspect expression signal and said alternate expression signal.

11. Apparatus according to claim 9, having the further improvement whereby the comparison means includes means for detecting a disparity type representative of transposition of character order, with respect to said alternate expression signal, within said suspect expression signal.

12. Apparatus according to claim 9, having the further improvement whereby the comparison means includes means for detecting a disparity type representative of deletion of characters, with respect to said alternate expression signal, from said suspect expression signal.

13. Apparatus according to claim 9, having the further improvement whereby the comparison means includes means for detecting a disparity type representative of unmatched characters, with respect to said alternate expression signal, within said suspect expression signal.

14. Apparatus according to claim 9, having the further improvement whereby the comparison means includes means for detecting a disparity type resulting from differences in the number of characters represented by said suspect expression signal and by said alternate expression signal.

15. Apparatus according to claim 9, having the further improvement whereby the comparison means includes means for detecting like characters represented in said suspect expression signal and said alternate expression signal.

16. Apparatus according to claim 10, having the further improvement whereby said comparison means further includes means for computing a disparity value as the summation of values indicated by each said significance signal.

17. Apparatus according to claim 9, having the improvement whereby said selection means includes auto-selection means coupled to said comparison means and responsive to a disparity signal representative of a value in a first numerical range for determining the alternate expression signal corresponding to a disparity signal to be exclusively substitutable for the suspect expression signal corresponding to said disparity signal and for producing a signal indicative thereof.

18. Apparatus according to claim 17, having the improvement whereby said auto-selection means includes means responsive to the number of characters represented by each of said suspect expression signal and said alternate expression signal being greater than a selected value, and for producing a signal indicative thereof.

19. Apparatus according to claim 17, having the further improvement whereby said selection means includes semi-automatic selection means coupled to said comparison means and responsive to a disparity signal representative of values in a second selected numerical range for selectively graphically displaying on said monitor one or more of said alternate expressions signals, said second selected numerical range having a lower value bound larger than a lower value bound of said first selected numerical range.

20. Apparatus according to claim 19, having the further improvement whereby said semi-automatic selection means includes means for accepting from said keyboard a signal indicative of an alternate expression signal, if any, to replace a corresponding suspect expression signal.

21. Apparatus according to claim 19, having the further improvement whereby said semi-automatic selection means includes means responsive to said disparity signal for ordering a display of alternate expression-representative signals on said monitor.

22. In a method for automated linguistic expression substitution on a digital data processor, the improvement wherein said digital data processor executes steps comprising

A. accepting into said digital data processor a suspect expression signal representative of a linguistic expression consisting of characters,

B. accepting into said digital data processor an alternate expression signal representative of a permissible linguistic expression consisting of characters,

C. comparing within said digital data processor said suspect expression signal with said alternate expression signal and producing a disparity signal numerically representative of differences between a spelling of the linguistic expression represented by said suspect expression signals and a spelling of the linguistic expression represented by said alternate expression signal, said comparing step including the step of producing said disparity signal to be numerically representative of the type and magnitude of differences between said suspect expression signal and said alternate expression signal, said comparing step further including the steps of responding to the detection of transposition, character deletion, unmatched character, and character length disparity types for producing a signal indicative of the numerically-weighted structural significance of that detected type, and

D. evaluating within said digital data processor a numerical value represented by said disparity signal for determining whether said alternate expression signal is substitutable for said suspect expression signal and for producing an output signal indicative thereof.

23. In a method for automated linguistic expression substitution on a digital data processor according to claim 22, the improvement in which said comparing step further comprises the step of detecting a disparity type representative of transposition of character order, with respect to said alternate expression signal, within said suspect expression signal.

24. In a method for automated linguistic expression substitution on a digital data processor according to claim 22, the improvement in which said comparing step further comprises the step of detecting a disparity type representative of deletion of characters, with respect to said alternate expression signal, from said suspect expression signal.

25. In a method for automated linguistic expression substitution on a digital data processor according to claim 22, the improvement in which said comparing step further comprises the step of detecting a disparity type representative of unmatched characters, with respect to said alternate expression signal, within said suspect expression signal.

26. In a method for automated linguistic expression substitution on a digital data processor according to claim 22, the improvement in which said comparing step further comprises the step of detecting a disparity type resulting from differences in the number of characters represented by said suspect expression signal and by said alternate expression signal.

27. In a method for automated linguistic expression substitution on a digital data processor according to claim 22, the improvement in which said comparing step further comprises the step of detecting like characters represented in said suspect expression signal and said alternate expression signal.

28. In a method for automated linguistic expression substitution on a digital data processor according to claim 22, the improvement in which said comparing step further comprises the step of computing a disparity value as the summation of values indicated by each said significance signal.

29. In a method for automated linguistic expression substitution on a digital data processor according to claim 22, the improvement in which the evaluating step further comprises the steps of responding to a disparity signal in a selected first range of disparity signal values for determining the alternate expression signal corresponding to a disparity signal to be exclusively substitutable for the suspect expression signal corresponding to said disparity signal and for producing a signal indicative thereof.

30. In a method for automated linguistic expression substitution on a digital data processor according to claim 29, the improvement in which the evaluating step further comprises the steps of responding to the number of characters represented by each of said suspect expression signal and said alternate expression signal being greater than a selected value and for producing a signal indicative thereof.

31. In a method for automated linguistic expression substitution on a digital data processor according to claim 29, the improvement in which the evaluating step further comprises the step of responding to a disparity signal value in a selected second range of disparity signal values and producing an output signal indicative thereof, said second disparity signal value range having a lower value bound larger than a lower value bound of said first disparity signal value range.