JPH05197752A - Machine translation system - Google Patents

Abstract

PURPOSE:To provide a machine translation system capable of efficiently obtaining a correct translated sentence. CONSTITUTION:This machine translation system provided with a function for receiving a word segmenting specification at the time of translating Japanese into another language and retranslating an original after its correction is provided with an input part 11 for inputting the original, a translation processing part 13 for executing translation processing by referring to a dictionary, an original morpheme conversion part 14 for displaying the original when it is not translated yet, or the original including word segmenting information after executing the translation processing on a screen, converting a morpheme expressed by 'KANA' (Japanese syllabary) in the original into a proper 'KANJI' (Chinese character) expression by referring to a corresponding table between headers in the dictionary and their 'KANA' expressions based upon a user's specification and specifying word segmentation, and an output part 15 for outputting a translated sentence obtained by the processing part 13.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a machine translation device,
In particular, the present invention relates to a system that can correct a morpheme in an original sentence and retranslate the corrected original sentence, or a machine translation device that can divide a morpheme in the original sentence (that is, specify word segmentation).

[0002]

2. Description of the Related Art In machine translation, a system sometimes makes a mistake in word segmentation or makes a segmentation different from the one intended by a user. This is especially likely to occur when translating an original text that has many Hiragana notations. Therefore, there has been proposed a machine translation device that allows a user to specify a word cut when the word cut is incorrect in the translation process or before translation. With this machine translation device, the original sentence containing the word cut information is displayed on the screen, and after the user consciously specifies the part to cut the word, the correct translation can be obtained by retranslating. Become.

[0003]

However, in such a conventional example, when the morpheme is in Hiragana notation even if the word is cut, it is often not registered in the dictionary in Hiragana notation, Since it is ambiguous even if it is registered, it was necessary to change the hiragana notation to kanji notation, or to register it in the dictionary with the heading of the hiragana notation, in addition to specifying the word break. In other words, it is an inefficient operation that the user often specifies the word segmentation for the original sentence with many hiragana notations, but it is difficult to obtain the correct translated sentence only with the word segmentation.

[0004]

SUMMARY OF THE INVENTION It is an object of the present invention to provide a machine translation device which is capable of improving the disadvantages of the conventional example and, in particular, being able to obtain a correct translated sentence by an efficiency opponent.

[0005]

According to the present invention, an original sentence is input in a machine translation system having a function of accepting a word cut specification when translating from Japanese into another language and correcting the original sentence and then retranslating it. Input unit and a translation processing unit that performs a translation process by referring to a dictionary. When the translation process is not performed, the original sentence is displayed, and after the translation process, the original sentence including the word segmentation information is displayed on the screen. A morpheme in the text that allows the user to specify a morpheme in Hiragana notation and convert it to an appropriate Kanji notation by referring to the correspondence table between the entry word and its Hiragana notation in the dictionary, according to the user's specification. The configuration is such that the conversion unit is provided and an output unit that outputs a translated sentence obtained by the translation processing unit is provided. This aims to achieve the above-mentioned object.

[0006]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS The present invention will be described below with reference to the drawings in accordance with embodiments. FIG. 1 is a configuration of the present invention, FIG. 2 is a configuration diagram showing a specific example of one embodiment of the present invention, and FIG. 3 is a diagram showing “headwords and their hiragana notation” in a dictionary referred to in one embodiment of the present invention. It is explanatory drawing which shows a part of "correspondence table." In the figure, reference numeral 11 is an input section, reference numeral 12 is a control section, reference numeral 13 is a translation processing section, reference numeral 14 is an original morphological conversion section, reference numeral 15 is an output section, reference numeral 21 is an input section, reference numeral 22 is a control section, reference numeral 23. Is a translation processing unit, reference numeral 24 is an original sentence morpheme conversion unit, and reference numeral 25 is an output unit.

The input unit 11 is for processing the input of an original sentence from a keyboard, for example. The control unit 12 has the following five functions. The original sentence is read from the input unit 11 and passed to the translation processing unit 13. The original sentence including the word segmentation information is read from the translation processing unit 13, and the original sentence morpheme converting unit 14 is read.
Hand over to. When untranslated, the original sentence is read from the input unit 11 and passed to the original sentence morpheme conversion unit 14. The corrected original sentence obtained by converting the morphemes in the original sentence is read from the original sentence morpheme converting unit 14 and passed to the translation processing unit 13. After the translation processing is completed by the translation processing unit 13, the translated text is taken out and passed to the output unit 15.

The translation processing unit 13 translates the original sentence provided from the input unit 11 and the corrected original sentence provided from the original sentence morpheme conversion unit 14 to generate a translated sentence in consideration of word segmentation. The source sentence morpheme conversion unit 14 displays the original sentence on the screen when it is not translated, and displays the original sentence including the word segmentation information after the translation process. In addition to performing the process of converting to, the original sentence morpheme is given the information for specifying the word segmentation. Further, the output unit 15 displays a translated sentence, which is a translation result of the translation processing unit 13, on a screen, for example.

Next, the operation of this embodiment will be described with reference to FIGS. Here, in Figure 2, once the original text has been translated,
The operation of correcting the original sentence including the word segmentation information in the original sentence morpheme conversion unit is shown. The correction specified by the user is performed as follows, for example. First, the control unit 22
Reads the original sentence "Wait until he comes." From the input unit 21 and passes it to the translation processing unit 23. Further, after the translation processing is completed, the control unit 22 extracts the original sentence including the word segmentation information from the translation processing unit 23 and sends it to the original sentence morpheme conversion unit 24. The original sentence morpheme conversion unit 24 displays the original sentence including the word segmentation information on the screen and receives the user's designation. The original sentence including the word segmentation information is displayed by inserting a space between the segmented morphemes in order to clarify the word segmentation. Therefore, if the translation process cuts words such as "he", "ga", "car", "de", and "wait", "He waits for car" is displayed as shown in FIG.

Now, the original sentence "Wait until he comes."
There are two possible interpretations: "Wait until he comes." And "Wait by car." Here, by adopting "wait until he comes.", The operation of converting the morpheme "Kuru" in Hiragana notation to the morpheme "Kuru" in Kanji notation is shown. When the user indicates on the screen the morpheme "Kuru" of the Hiragana notation in the original text with a mouse etc., the dictionary will search for the entry word "Kuru" in Hiragana reading and the words that match that condition will be selected as conversion candidates. Is displayed.

Here, as a method for searching for a headword,
Candidates can be easily searched by referring to the correspondence table of headwords and their hiragana notation in the dictionary shown in FIG. From the table of FIG. 3, three types of “come”, “roll”, and “cure” are displayed as candidates. When the user selects an appropriate headword (“come”) from the candidates by some operation, the morpheme is automatically converted into kanji notation, and the correction result of the original text is “He comes Wait until. " Further, the morpheme "come" designated by the mouse or the like is designated as a word break. Subsequently, the control unit 22 takes out the corrected original sentence from the original sentence morpheme conversion unit 24 and receives the word segmentation designation.
Hand over to. The translation processing unit 23 generates a translated sentence so as to reflect the morpheme converted by the original sentence morpheme converting unit 24 and the word segmentation designation. Finally, the control unit 22 uses the translation processing unit 23.
The translated text is taken out from and passed to the output unit 25.

Further, in the case of adopting "He waits in the car.", The user indicates on the screen the morpheme "car" of the Hiragana notation in the original text. Then, the table of FIG. 3 is searched for a headword in the dictionary, which is “car” in hiragana reading, and the word “car” that meets the condition is displayed as a conversion candidate. Subsequent operations are the same as when adopting.

[0013]

As described above, according to the present invention, it is possible to accept a word cut specification when translating from Japanese into another language, and to retranslate after correcting the original sentence. For the morpheme of Hiragana notation in the original sentence, as an example of the search method, by referring to the correspondence table of the entry word in the dictionary and its Hiragana notation, it is used from the corresponding morpheme of Kanji notation Converted to the appropriate Kanji notation selected by the person
Further, by retranslating the morpheme with the word segmentation designated, it is possible to provide an excellent machine translation device that does not exist in the related art in which a correct translated sentence can be efficiently obtained.

[Brief description of drawings]

FIG. 1 is a configuration diagram showing an embodiment of the present invention.

FIG. 2 is an explanatory diagram showing the configuration and operation of one embodiment of the present invention.

FIG. 3 is an explanatory diagram showing a part of a correspondence table of headwords and their hiragana notation in a dictionary referred to in an embodiment of the present invention.

[Explanation of symbols]

 11,21 Input unit 12,22 Control unit 13,23 Translation processing unit 14,24 Original sentence morpheme conversion unit 15,25 Output unit

Claims (1)

[Claims] 1. In a machine translation system having a function of accepting a word cut specification when translating from Japanese into another language and retranslating after correcting the original sentence, refer to an input unit for inputting the original sentence and a dictionary. It has a translation processing unit that performs translation processing, and displays the original sentence when it is not translated and the original sentence that includes the word break information after the translation process, respectively, for the morphemes in Hiragana notation in the original sentence. According to the user's designation, the corresponding table of the headword in the dictionary and its hiragana notation is referred to and converted into an appropriate Kanji notation, and the original sentence morpheme conversion unit capable of designating the word cut is equipped and A machine translation device, comprising: an output unit that outputs a translated text obtained by the translation processing unit.