Pre-edit support method and apparatus

• US 5,396,419 A
• Filed: 09/08/1992
• Issued: 03/07/1995
• Est. Priority Date: 09/07/1991
• Status: Expired due to Fees

• Current Assignee: Hitachi, Ltd.
• Original Assignee: Hitachi, Ltd.
• Inventors: Morimoto, Yasutsugu
• Primary Examiner(s): Hayes, Gail O.
• Assistant Examiner(s): CHUNG TRANS, XUONG MY
• Application Number:
  US07/942,030
• Time in Patent Office: 910 Days
• Field of Search: 364/419, 364/419.02, 364/419.08, 364/419.04
• US Class Current: 704/2
• CPC Class Codes: G06F 40/216 using statistical methods
• CPC Class Codes: G06F 40/268 Morphological analysis
• CPC Class Codes: G06F 40/44 Statistical methods, e.g. probability models

Abstract

A pre-edit support method for a machine translation system has an improved accuracy of parsing and an improved efficiency of pre-edit support. In a machine translation system translating a source language into a target language, a source language text is read and divided into words. Word sequences occurring a plurality of times are extracted, of the extracted word sequences the word sequences having a high occurrence frequency or a large number/high ratio of words with multiple parts of speech, are extracted as complex word candidates and displayed to a user, of the complex word candidates the complex word candidates the user designates are used as complex words to be translated integrally, and pre-edit symbols are added to the integral complex words as translation support information.

Claims

1. In a machine translation system translating a source language into a target language, a translation support method comprising the steps of:

reading a source language text;

dividing said text into words;

extracting, from sequential said words from said dividing step, sequential word groups which occur a plurality of times throughout said text;

displaying ones of extracted said sequential word groups satisfying a predetermined occurrence frequency, as complex word candidates; and

allowing user selection of at least one complex word candidate of said complex word candidates, to be treated as a complex word which is to be translated as an integral complex word.

2. A translation support method as claimed in claim 1, further comprising the step of:

inserting translation support information in a display of said text by providing a predetermined indication at positions where said at least one complex word candidate occurs.

3. A translation support method as claimed in claim 1, further comprising the steps of:

displaying at least one translation of each word of said at least one complex word candidate;

selecting a correct translation of each word from displayed said at least one translation; and

adding a translation of said at least one complex word candidate to a dictionary; and

providing a correction of a displayed translation.

4. A translation support method as claimed in claim 1, comprising the further step of;

setting predetermined inclusion relations defining conditions for displayed said complex word candidates, wherein displayed said complex word candidates satisfy a predetermined condition based on said predetermined inclusion relations in addition to said predetermined occurrence frequency, and wherein said predetermined condition requires a sequential word group of said sequential word groups to be one of a root complex word candidate which has no parent complex word candidate or a child complex word candidate which occurs independently of the parent complex word.

5. A translation support method as claimed in claim 1, comprising the further step of:

selecting a display mode for displaying ones of said complex word candidates, wherein said display mode includes the modes of;

at least one of a candidate list display mode and an in-text emphasis display mode.

6. In a machine translation system translating a source language into a target language, an automatic translation method comprising the steps of:

reading a source language text;

dividing said text into words;

extracting, from sequential said words, at least one sequential word group occurring a predetermined number of times throughout said text as a complex word candidate to be presented to a user for possible selection as a complex word; and

translating a selected said complex word as an integrated complex word, into said target language.

7. In a machine translation system translating a source language into a target language, a translation support apparatus comprising:

a reading means for reading a source language text;

a dividing means for dividing said text into words;

an extracting means for extracting, from sequential said words from said dividing means, sequential word groups occurring a plurality of times throughout said text;

a candidate means for displaying extracted said sequential word groups satisfying a predetermined occurrence frequency as complex word candidates; and

a selecting means for user selecting, from displayed said sequential word groups from said candidate means, a complex word which is to be translated as an integral complex word.

8. A translation support apparatus as claimed in claim 7, further comprising:

a complex word indication means for inserting translation support information in a display of said text by providing a predetermined indication at positions where a selected said complex word occurs.

9. In a machine translation system translating a source language into a target language, an automatic translation apparatus comprising:

reading means for reading a source language text;

a dividing means for dividing said text into words;

an extracting means for extracting, from sequential said words from said dividing means, sequential word groups occurring a predetermined number of times throughout said text as complex word candidates to be presented to a user for possible selection as a complex word; and

a translating means for translating a selected said complex word as an integrated complex word, into said target language.

10. In a machine translation system translating a source language into a target language, a translation support method comprising the steps of:

reading a source language text;

dividing said text into words;

setting predetermined inclusion relations defining conditions required for designation of sequential word groups as complex word candidates;

extracting sequential word groups, from sequential said words from said dividing step, which satisfy said predetermined inclusion relations and which occur a plurality of times throughout said text as said complex word candidates;

storing an occurrence position of each extracted complex word candidate of said complex word candidates in said text;

comparing each occurrence position of each of said extracted complex word candidates with occurrence positions of parent complex word candidates of said complex word candidates, to thereby extract a complex word candidate having an occurrence position at which no parent complex word candidate occurs;

displaying any said complex word candidate of said complex word candidates satisfying a predetermined occurrence frequency; and

selecting at least one complex word candidate from displayed said complex word candidates which is to be treated as a complex word.

11. In a machine translation system translating a source language into a target language, a translation support method comprising the steps of:

reading a source language text;

dividing said text into words;

extracting, from sequential said words from said dividing step, sequential word groups which occur a plurality of times throughout said text, wherein said step of extracting word sequences takes into consideration, in a determination of whether a word sequence is to be extracted, at least one of;

(a) whether said word sequence includes words which can be categorized into alternative parts of speech, and (b) a ratio of words which can be categorized into alternative parts of speech to a total number of words on said word sequence;

displaying ones of extracted said sequential word groups satisfying a predetermined occurrence frequency, as complex word candidates; and

allowing user selection of at least one complex word candidate of said complex word candidates, to be treated as a complex word which is to be translated as an integral complex word.