JPH08202720A - Machine translation system - Google Patents

Abstract

PURPOSE: To comprehend a translated sentence more easily for a user by outputting a specified word while keeping it in an original sentence as it is based on the information of words not to be translated applied to a non- translation information holding part. CONSTITUTION: When designating the word not to be translated, data showing the word not to be translated are stored in a non-translation information holding part 7. Next, one sentence is read by an input part 1. Afterwards, analystic processing is performed by an analystic processing part 2 based on a dictionary 5 and grammatical rules 6 and then, translation generating processing is performed by a translation generating processing part 3 based on the dictionary 5 and the grammatical rules 6. Next, it is judged whether the data are stored in the non-translation information holding part 7 or not before the translation generated result is outputted by an output part 4 and when the data are not stored, the translated word is outputted. When the data are stored, it is judged whether the kind of the word to be outputted is matched with the data in the non-translation information holding part 7 or not and when they are not matched, the translated word is outputted. When they are matched, the original sentence is outputted as it is.

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 that makes no translation depending on the type of word.

[0002]

2. Description of the Related Art In a conventional machine translation device, all input sentences input in a first language are converted (translated) into a second language and output, so that they are used only in a specific field in the input sentence. When there are technical terms such as those, when translating those terms, it was rather difficult to read the meaning of the translated text.

For example, the compound noun phrase "gearbox modu
le "with the individual nouns that make it up," gearbox "," module "
There was a problem that it was divided and translated and was output as "gearbox component", for example.

With respect to such a problem, Japanese Patent Laid-Open No. 3-22
2065 discloses a machine translation device capable of outputting a part having a specific grammatical structure in an input sentence as an original sentence. For example, a noun phrase in which two or more nouns are continuous or a noun phrase consisting of one or more adjectives in which nouns are continuous is targeted for non-translation.

As a concrete example, when the sentence "The system requires a gearbox module." Is input as an input sentence, the analysis result is "The system requires a gearbox module."

[0006] For example, when a noun phrase consisting of an article + two or more consecutive nouns is a target of no conversion, "a gearb
According to this technology, the technical terms with the structure of compound noun phrases are divided into the individual nouns that compose them, and a difficult-to-understand translated sentence is output. You can solve the problem of doing.

[0007]

However, in the conventional machine translation device, a part that does not have a specific grammatical structure is
All the words are translated into the target language, so even if the input sentence contains a word that is easier for the user to understand as it is, the word will be translated and the meaning of the translated sentence will be harder to read. There were cases.

It is an object of the present invention to provide a machine translation device in which a user can specify that a word retrieved from a specific dictionary should be output as the original sentence, and a word belonging to a specific part of speech is output as the original sentence. The purpose is to provide a machine translation device that can be specified by the user.

[0009]

In order to solve the above problems, an input unit for inputting a sentence in a first language, a dictionary for analyzing an input sentence in the first language input from the input unit, and a grammatical rule An analysis processing unit for analyzing using the sentence, a conversion generation processing unit for converting the sentence analyzed by the analysis processing unit into a second language for translation, and an output unit for outputting the sentence converted by the conversion generation processing unit. In the machine translation device, whether the untranslated information holding unit that stores the information of the word that should be untranslated and the word of the sentence of the second language output at the output unit are the words stored in the untranslated information holding unit Make a decision,
For a specific word that should be untranslated, a non-translation information providing unit that outputs the corresponding word as the first language is provided.

A dictionary in which the type of dictionary to be untranslated is specified in the untranslated information holding unit, and the word of the sentence of the second language output in the output unit is specified in the untranslated information holding unit. It is determined whether the word is a word retrieved from, and if it is a word in a dictionary that should be non-translated, a non-translation information addition unit that outputs the corresponding word as the first language is provided.

Further, the part of speech to be untranslated is specified in the untranslated information holding unit, and the word of the sentence of the second language output in the output unit is the word of the part of speech specified in the untranslated information holding unit. If it is a part of speech that should be untranslated, the problem is solved by providing an untranslated information adding unit that outputs the corresponding word in the first language as it is.

[0012]

[Action]

[Operation of Claim 1] In the machine translation apparatus configured as described above, the analysis processing unit and the conversion generation processing unit output the result converted into the sentence of the target language based on the dictionary and the grammatical rule at the output unit. When output, the untranslated information adding unit outputs a specific word as the original text based on the information of the word that should be untranslated added to the untranslated information holding unit. Will be easier to understand.

[Operation of Claim 2] In the machine translation apparatus configured as described above, in the analysis processing unit and the conversion generation processing unit,
When the result converted into the sentence of the target language based on the dictionary and the grammatical rule is output at the output unit, the untranslated information adding unit previously sets the untranslated dictionary of the dictionary to be added to the untranslated information holding unit. Based on the information, the word retrieved from the specific dictionary is output as the original sentence, which makes it easier for the user to understand the translated sentence.

[Operation of Claim 3] In the machine translation apparatus configured as described above, in the analysis processing unit and the conversion generation processing unit,
When the result converted into the sentence of the target language based on the dictionary and the grammatical rule is output at the output unit, the untranslated information adding unit pre-defines the part of speech that has been added to the untranslated information holding unit. Based on the information, the word of a specific part of speech is output as the original sentence, which makes it easier for the user to understand the translated sentence.

[0015]

An embodiment will be described with reference to the drawings.

FIG. 1 is a block diagram showing a main part of a machine translation device according to the present invention. With reference to FIG. 1, the machine translation apparatus uses an input unit 1 for reading a sentence in an input language desired to be machine translated, an analysis for analyzing a sentence in the input language using a dictionary 5 and a grammar rule 6. A processing unit 2, a conversion generation processing unit 3 for converting a sentence in a language input by using a dictionary 5 and a grammar rule 6 into a target language, and generating the same, a non-translation information holding unit 7, a non-translation information addition unit 8, and a translation. And an output unit 4 for outputting the resulting target language. In the input unit 1, not only file information but also a text may be directly input from a keyboard.

FIG. 2 is a configuration diagram showing the configuration of the untranslated information holding unit 7 shown in FIG. The untranslated information holding unit 7 includes at least an untranslated dictionary information storage area 71 and an untranslated part-of-speech information storage area 72. In each area, information of a word to be non-translated, which is added by the non-translation information adding unit 8, is held.

FIG. 3 is a flow chart showing the flow of processing in the machine translation apparatus shown in FIG. Referring to FIG. 3, first, it is determined whether a specific type of word to be untranslated is designated (step S11), and if a word to be untranslated is not designated, the process proceeds to step S13. If the word to be untranslated is specified, the data indicating the word to be untranslated is stored in the untranslated information holding unit 7 (step S12). Next, one sentence is read by the input unit 1 (step S13). Next, the analysis processing unit 2 performs analysis processing based on the dictionary 5 and the grammar rule 6 (step S14), and then the conversion generation processing unit 3 performs conversion generation processing based on the dictionary 5 and the grammar rule 6 (step S15). . Next, before outputting the conversion generation result at the output unit 4, it is determined whether or not the data is stored in the non-translation information holding unit 7 (step S1).
6) If the data is not stored, the translated word is output (step S19). If the data is stored, it is determined whether the type of the word to be output and the data in the untranslated information holding unit 7 match (step S17). If they do not match, the translated word is output (step S19). If they match, the original text is output (step S18).

[Embodiment of Claim 2] FIG. 4 is a flow chart showing the flow of processing when a dictionary to be translated is designated. Referring to FIG. 4, first, it is determined whether or not a dictionary to be translated is designated (step S21). If a dictionary to be translated is not designated, the process proceeds to step S23. If the untranslated dictionary is designated, data indicating the untranslated dictionary is stored in the untranslated dictionary information storage area 71 (step S22). That is, for each type of dictionary, for example, the number assigned in advance is stored, such as "1" for the basic word dictionary, "2" for the technical term dictionary, and "3" for the user dictionary. Next, one sentence is read by the input unit 1 (step S23). Next, the analysis processing unit 2 performs analysis processing based on the dictionary 5 and grammar rule 6 (step S24), and then the conversion generation processing unit 3 performs conversion generation processing based on the dictionary 5 and grammar rule 6 (step S24). S2
5). Next, before outputting the conversion generation result by the output unit 4, it is judged whether or not the data is stored in the untranslated dictionary information storage area 71 (step S26), and if the data is not stored, the translated word is output. (Step S2
9). If the data is stored, it is judged whether or not the type of the dictionary of the word to be output and the data of the untranslated dictionary information storage area 71 match (step S27), and if they do not match, the translated word is output (step S27). S29). If they match, the original text is output (step S28).

As a concrete example, "Check the memory cel
Explain the case of translating the English sentence "l." into Japanese. This English sentence consists of three words, "check", "the" and "memory cell". Of these three words, "memory cell"
Only two terms are taken from the technical term dictionary, and the other two words are taken from the basic term dictionary. Then, it is assumed that each translated word is “confirm”, “(no translated word)”, and “memory cell”. At this time, the situation of the dictionary of each word is as shown in FIG.

If the technical term dictionary is not designated as the untranslated dictionary, the steps S23, S24, S25 and S26 are performed, and then the step S29.
At, the translated word corresponding to each word is output, and the translated text becomes "Check memory cell". Here, step S
24, the structure obtained as a result of step S25 is as shown in FIG.

Next, if the technical term dictionary is designated as the untranslated dictionary, in step S22, the data "2" indicating the untranslated dictionary, that is, the technical term dictionary, is stored in the untranslated dictionary information storage area 71. To store. The situation of the untranslated dictionary information storage area 71 at this time is as shown in FIG.
Then, step S23, step S24, step S
After performing the processing of 25, when it is judged in step S26 whether or not the data is stored in the untranslated dictionary information storage area 71, the data is stored, so the processing proceeds to step S27 and the dictionary type of the word to be output. And whether the types of dictionaries indicated by the data “2” in the untranslated dictionary information storage area 71 match are checked for each word to be output. First, "chec
Since "k" is not a word drawn from the technical term dictionary, the translated word is output in step S29. Next, since "the" is not a word drawn from the technical term dictionary, the translated word is output in step S29. Next, since "memory cell" is a word drawn from the technical term dictionary, it is output as the original text in step S28. As a result, the translated text is "to check the memory cell", and a translated text that is easier to understand can be obtained.

[Embodiment of claim 3] FIG. 8 is a flow chart showing a flow of processing when a part of speech desired to be untranslated is designated. Referring to FIG. 8, first, it is determined whether or not the part of speech to be untranslated is designated (step S31). If the part of speech to be untranslated is not designated, the process proceeds to step S33. If the part of speech to be untranslated is designated, data indicating the part of speech to be untranslated is stored in the untranslated part of speech information storage area 72 (step S32). That is, for each type of part-of-speech, a number assigned in advance is stored, such as "1" for nouns, "2" for pronouns, and "3" for proper nouns.
Next, one sentence is read by the input unit 1 (step S3
3). Next, the analysis processing unit 2 performs analysis processing based on the dictionary 5 and the grammar rule 6 (step S34), and then the conversion generation processing unit 3 performs conversion generation processing based on the dictionary 5 and the grammar rule 6 (step S15). ). Next, before outputting the conversion generation result by the output unit 4, it is judged whether or not the data is stored in the untranslated part-of-speech information storage area 72 (step S36), and if the data is not stored, the translated word is output. (Step S39). If the data is stored, it is determined whether the type of part of speech of the output word and the data in the untranslated part-of-speech information storage area 72 match (step S37), and if they do not match, the translated word is output (step S39). ). If they match, the original text is output (step S38).

As a concrete example, a case where a Japanese sentence "Honda is different from Honda" is translated into English will be described. Both “Honda” and “Honda” are proper nouns, and both are “Hond”.
It shall have the translation "a". At this time, the situation of the dictionary of each word is as shown in FIG.

If no part of speech to be translated is not designated, step S33, step S14, step S3
5, through step S36, the translated word corresponding to each word is output in step S39, and the translated text is "Honda is different
from Honda. ” Here, step S2
4, the structure obtained as a result of step S25 is as shown in FIG.

Next, if a proper noun is designated for the part of speech to be untranslated, data "3" representing the part of speech to be untranslated, that is, the proper noun is stored in the untranslated part-of-speech information storage area 72 in step S32. It The situation of the untranslated dictionary information storage area 71 at this time is as shown in FIG. Next, through step S33, step S34, step S35, and step S36, for each word output in step S37, the type of part of speech and the untranslated part-of-speech information storage area 7
It is determined whether or not the types of parts of speech indicated by the data of 2 match. Since both "Honda" and "Honda" are proper nouns, the original text is output in step S38. As a result, the translated sentence becomes "Honda is different from Honda.", And a translated sentence that is easier to understand can be obtained.

[0027]

By providing the non-translation information holding unit in the machine translation device, the vocabulary retrieved from the specific dictionary in the input sentence can be output as the original sentence, which is more comprehensible than when unnecessary translation is performed. Easy to output translation results.

Further, by providing the non-translation information holding unit in the machine translation device, the vocabulary of a specific part of speech in the input sentence can be output as the original sentence, and it is easier to understand than when unnecessary translation is performed. The translation result can be output.

[Brief description of drawings]

FIG. 1 is a block diagram showing a main part of a machine translation device of the present invention.

FIG. 2 is a configuration diagram showing a configuration of a non-translation information holding unit 7 shown in FIG.

FIG. 3 is a flowchart showing a flow of general processing in the case of designating a type of a word that is not translated in the machine translation device shown in FIG.

FIG. 4 is a flowchart showing a flow of processing in the case of designating a dictionary type that is not translated in the machine translation device shown in FIG.

FIG. 5 is a diagram showing a situation of a dictionary assumed in a specific example of the processing shown in FIG.

FIG. 6 is a diagram showing a display example of an internal structure in a specific example of the processing shown in FIG.

7 is a configuration diagram showing a situation in which data is stored in the untranslated information holding unit 7 shown in FIG.

8 is a flowchart showing the flow of processing when the type of part of speech to be untranslated is designated in the machine translation device shown in FIG.

FIG. 9 is a diagram showing a situation of a dictionary assumed in a specific example of the processing shown in FIG.

10 is a diagram showing a display example of the internal structure in the specific example of the processing shown in FIG.

11 is a configuration diagram showing a situation where data is stored in the non-translation information holding unit 7 shown in FIG.

[Description of Signs] 1: Input unit 2: Analysis processing unit 3: Conversion generation processing unit 4: Output unit 5: Dictionary 6: Grammar rule 7: Non-translation information holding unit 8: Non-translation information addition unit

Claims (3)

[Claims] 1. An input unit for inputting a sentence in a first language, an analysis processing unit for analyzing an input sentence in the first language input from the input unit using a dictionary for analysis and grammar rules, and the analysis processing. In a machine translation device that includes a conversion generation processing unit that converts a sentence analyzed by a unit into a second language for translation, and an output unit that outputs the sentence converted by the conversion generation processing unit, a word to be untranslated And a non-translation information storage unit that stores the information of the second sentence, and whether or not the word of the sentence of the second language output at the output unit is a word stored in the non-translation information storage unit is determined from the information from the dictionary unit If it is a specific word that should be
A machine translation device comprising a non-translation information providing unit that outputs a corresponding word as a first language. 2. A dictionary in which the type of dictionary to be untranslated is specified in the untranslated information holding unit, and the word of the sentence of the second language output in the output unit is specified in the untranslated information holding unit. The non-translation information adding section is provided for determining whether the word is a word retrieved from the dictionary and outputting the corresponding word in the first language as it is if the word is a dictionary word that should be untranslated.
The machine translation device described. 3. The part of speech to be untranslated is specified in the untranslated information holding unit, and the word of the sentence in the second language output in the output unit is the word of the part of speech specified in the untranslated information holding unit. 2. The machine translation device according to claim 1, further comprising a non-translation information adding unit that outputs the corresponding word in the first language if it is a part of speech that should be determined to be non-translation.