JPH04367071A - Machine translation device - Google Patents

Abstract

PURPOSE:To decrease the dictionary preparing work and to reduce the dictionary information by providing a Japanese language analyzing part and analyzing the translated word in an English-Japanese dictionary. CONSTITUTION:An English language analyzing part 5 divides an input sentence into each word, refers to an English-Japanese word dictionary 3 for the divided English word and takes out speech part and the translated word concerning each word. An English-Japanese converting part 6 receives the output of the English language analyzing part 5, and while it refers to a translation rule dictionary 4, converts to the Japanese language mimic to the syntax of an English grammar from the syntax of the English language, and a Japanese language analyzing part 10 refers to the English-Japanese word dictionary 3 and the translation rule dictionary 4 and analyzes it. Further, the result of the processing up to now and the translation rule dictionary 4 are referred to and a Japanese language preparing part 7 forms the Japanese language. At this time, a translation performing part 8 writes the result into an original sentence and translated sentence pair control table 2, and the translation performing part 8 refers to the original sentence and translated sentence pair control table 2 and displays 9 the translated sentence. Then, since it is not necessary to store the information equivalent to the analyzed result concerning the translated word in the dictionary, the dictionary preparing work can be decreased and the dictionary information can be reduced.

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 for inputting a language and converting it into another language.

0002

[Prior Art] In recent years, with machine translation devices that have been widely used as devices for translating from one language to another, translators have been involved in analyzing sentences during the translation process. .

FIG. 4 is a functional block diagram of a conventional machine translation device. In the figure, 1 is an input means having a means for inputting the original text and instructing translation, 2 is an original-translation pair management table that stores the original text inputted from the input means 1 and the translation result in correspondence, and 3 is an analysis of English. 4 is a dictionary that stores grammatical information about English words necessary for reading English words, translations of these English words, and Japanese words and their English grammatical information; 4. A translation rule dictionary 5 stores syntax correspondence information and grammars translated from English grammar, and 5 analyzes the English sentence input from the input means 1 with reference to the English-Japanese word dictionary 3 and the translation rule dictionary 4. An English analysis unit, 6 converts the output of the English analysis unit 5 into an input syntax with reference to the translation rule dictionary 4, and provides a translation for each English word; 7, the English-Japanese conversion unit; A Japanese generation unit generates a Japanese sentence by adding particles and auxiliary verbs to the output of unit 6 or by utilizing the endings of preverbs; 8 is the original text stored in the source/translation pair management table 2; The translation execution unit 9 is a display means that displays the original text and the translated text stored in the original text/translation pair management table 2 in association with each other.

FIG. 2 is a block diagram showing the hardware configuration of a machine translation device. In the figure, 11 is the keyboard, 12
1 is a central processing unit (CPU), 13 is a cathode ray tube display (CRT), 14 is ROM, and 15 is RAM.

The input means 1 is provided by a keyboard 11, the original-translation pair management table 2 is provided by a RAM 15, the English-Japanese word dictionary 3 and the translation rule dictionary 4 are provided by a ROM 14, an English analysis section 5, an English-Japanese conversion section 6, and a Japanese generation section. 7. The translation execution section 8 is realized by the CPU 12 executing a program stored in the ROM 14 while exchanging data with the ROM 14 and the RAM 15, and the display means 9 is realized by the CRT 13.

FIG. 3 is a flowchart showing the operation of the machine translation device.

[0007] The translator inputs an English sentence from the input means 1 (
S1). The input English sentence is stored in the original-translation sentence pair management table 2 by the translation execution unit 8.

[0008]Here, explanation will be given assuming that the following example sentence "He was not popular." has been input.

First, the English analysis unit 5 divides the input sentence into words. In the above example, “He”, “was”, “no”
t”, “popular”, “. ''. Next, the English-Japanese word dictionary 3 is referred to for these divided English words, and the part of speech and translation for each word are extracted. Table 1 shows the ``information about the words''. Furthermore, the English analysis unit 5 refers to the English grammar translation rule dictionary 4 and recognizes the syntax of the input sentence in English grammar (S2). In the above example, the part “be not” is the predicate V,
It recognizes that "he" is the subject S, that "popular" is the complement C, and that the syntax as a whole is "SVC". Table 2 shows the ``English grammar rules used for analysis'' of this sentence.

The English-Japanese conversion unit 6 receives the output of the English sentence analysis unit 5, and converts the English sentence into a Japanese sentence that follows the English grammar while referring to the translation rule dictionary 4 (S3). In the above example, the rule is adopted that ``If the syntax is SVC and the predicate is composed of the verb be, C must be translated as a predicate in Japanese.''

[0011] The translator refers to the English-Japanese word dictionary 3 and translation rule dictionary 4 for the translated word information extracted in (S1) (S4).
), ``he'' is a noun, ``is'' is an auxiliary verb, ``nai'' is an auxiliary adjective, and ``popular'' is understood. Therefore, the translator inputs into the translation rule dictionary 4 from the keyboard 11 the information that the subject is composed of the noun ``popular'', each particle ``ga'', and the auxiliary verb ``aru'' in the form of subject + predicate. Must.

[0012] Results of processing so far and translation rule dictionary 4
The Japanese generation unit 7 generates Japanese by referring to (S5)
. In other words, we can see that ``the subject is ``he'' and the predicates are ``popular'' and ``not'', so the tense must be synthesized as the past tense.'' Furthermore, from the result of (S4), "popular" is a combination of the noun "popular" and the case particle "ga".
We know that the word is composed of a subject + predicate with the auxiliary verb ``there is'', so if we combine ``popular'' and ``not'' into the past tense, we get ``not popular''. Table 4 shows the rules adopted at this time. Based on the above results, the input English sentence is translated as "he was not popular." At this time, the translation execution unit 8 writes the result into the original/translated text pair management table 2.

The translation execution unit 8 refers to the original translation pair management table 2 and displays "He was not popular" on the display means 9 (S6). Table 5 shows the "state of the display screen" at this time.

[0014]

[Problem to be solved by the invention] The process up until now has been to analyze the Japanese language with the help of a translator, and to analyze the Japanese language using the past tense, since the subject is ``he'' and the predicates are ``popular'' and ``nai.'' It is assumed that ``popular'' is composed of the noun ``popular'', the case particle ``ga'', and the auxiliary verb ``aru'' in the form of subject + predicate. , this syntax was stored in the translation rule unit 4. In other words, each translator had to participate in the translation work.

The present invention has been made in view of the above-mentioned conventional problems, and an object of the present invention is to provide a means for transferring as much of the work of Japanese language analysis to a machine translation device as possible.

[0016]

[Means for Solving the Problems] The present invention provides an input means having means for inputting an original text and instructing translation, and an original-translation correspondence management table that stores the original text inputted from the input means in association with the translation result. , grammatical information about English words necessary to analyze English, translations for these English words, an English-Japanese word dictionary that stores Japanese words and their English grammatical information, and grammatical and grammatical information about English words. A translation rule dictionary that stores correspondence information between English and Japanese syntax and grammars translated from English grammar, and the English sentence inputted from the input means are analyzed by referring to the English-Japanese word dictionary and the translation rule dictionary. an English analysis section, an English-Japanese conversion section that converts the output of the English analysis section into an input syntax by referring to the translation rule dictionary, and provides a translation for each English word; a Japanese generation unit that generates a Japanese sentence by adding particles or auxiliary verbs or utilizing the endings of predicates; and a translation execution unit that translates the original text stored in the source/translation pair management table. and display means for displaying the original text and the translated text stored in the original text/translation pair management table in association with each other. It is characterized by having a Japanese language analysis section that analyzes and makes decisions based on Japanese grammar.

[0017]

[Operation] According to the present invention, by adding an analysis means using Japanese grammar to the flow of operations of a conventional machine translation device, the work of creating a dictionary is reduced, and the capacity of the dictionary is also reduced.

[0018]

DESCRIPTION OF THE PREFERRED EMBODIMENTS FIG. 1 is a functional block diagram of a machine translation apparatus according to an embodiment of the present invention. In the figure, 10 is a Japanese language analysis section. Other numerals 1 to 9 are the same as the conventional one shown in FIG. In addition, the hardware configuration is the same as in FIG. 2,
The operation will be explained using FIG. 3 mentioned above.

Translation inputs English text from input means 1 (S
1). The input English sentence is stored in the original-translation sentence pair management table 2 by the translation execution unit 8.

[0020] Here, explanation will be given assuming that the following example sentence "He was not popular." has been input.

First, the English analysis unit 5 divides the input sentence into words. In the above example, “He”, “was”, “no”
t”, “popular”, “. ”.Next, refer to the English-Japanese word dictionary 3 for these divided English words,
Extract the part of speech and translation for each word. Information for each word is shown in Table 1. Furthermore, the English analysis unit 5 refers to the English grammar translation rule dictionary 4 and recognizes what kind of English grammar structure the input sentence has (S2). In the above example, the part “be not” is the predicate V, and “he
” is the subject S, “popular” is the complement C, and the overall structure is “SVC”. Table 2 shows the rules of English grammar used to analyze this sentence.

The English-Japanese converter 6 receives the output of the English analyzer 5, and converts the English syntax into a Japanese sentence modeled on the English grammar while referring to the translation rule dictionary 4 (S3). In the above example, the rule is adopted that ``If the syntax is SVC and the predicate is composed of the verb be, C must be translated as a predicate in Japanese.''

The Japanese language analysis unit 10, which is the main part of the present invention, analyzes the translated information extracted in (S1) with reference to the English-Japanese word dictionary 3 and the translation rule dictionary 4 (S4). Here, "he" is a noun, "deru" is an auxiliary verb, "nai" is an auxiliary adjective, and "popular" is the noun "popular" and "ga"
The subject + by the case particle and the auxiliary verb ``aru''
The Japanese language analysis unit 10 recognizes that the word is composed of a predicate.
recognizes. Table 3 shows the "translated words and analysis results for the translated words" at this time.

[0024] Results of processing so far and translation rule dictionary 4
The Japanese language generation unit 7 forms Japanese with reference to (S5)
. In other words, we can see that ``the subject is ``he'' and the predicates are ``popular'' and ``not'', so the tense must be synthesized as the past tense.'' Furthermore, from the result of step (S4),
We know that "popular" is a word composed of the noun "popular", the case particle "ga", and the auxiliary verb "aru" in the form of subject + predicate, so "popular" If we combine the past tense from ``not'' and ``not'', we get ``it was not popular.'' Table 4 shows the rules adopted at this time.
Shown below. From the above results, the translation for the input English sentence is “
He was not popular." At this time, the translation execution unit 8 writes the result into the original/translated text pair management table 2.

[0025] The translation execution unit 8 uses the source/translation pair management table 2
, and displays "He was not popular" on the display means 9 (S6). Table 5 shows the “display screen appearance” at this time.
Shown below.

Although the present invention has been described with respect to English-Japanese translation, it goes without saying that it can be applied to other languages as well.

[0027]

[Table 1]

[0028]

[Table 2]

[0029]

[Table 3]

[0030]

[Table 4]

[0031]

[Table 5]

[0032]

[Effects of the Invention] As explained above, the machine translation device of the present invention includes a Japanese language analysis section to analyze translated words in an English-Japanese dictionary, so the dictionary stores information corresponding to the analysis results regarding translated words. There's no need to. Therefore, the dictionary creation work can be reduced and the dictionary information situation can also be reduced.

[Brief explanation of drawings]

FIG. 1 is a functional block diagram of a machine translation device according to an embodiment of the present invention.

FIG. 2 is a block diagram showing the hardware configuration of the present invention and a conventional machine translation device.

FIG. 3 is a flowchart showing the operation of the present invention and a conventional machine translation device.

FIG. 4 is a functional block diagram of a conventional machine translation device.

[Explanation of symbols]

1... Input means, 2... Original text translation pair management table,
3...English-Japanese word dictionary, 4...Translation rule dictionary, 5...
English analysis unit, 6...English-Japanese conversion unit, 7...Japanese generation unit, 8...translation execution unit, 9...display means, 10...Japanese analysis unit.

Claims (2)

[Claims] 1. An input means having a means for inputting an original text and instructing translation; an original translation pair management table storing the original text inputted from the input means and the translation result in association with each other;
Grammatical information about English words needed to parse English,
It also stores translations for these English words, an English-Japanese word dictionary that stores Japanese words and their English grammatical information, and English grammar, correspondence information between English and Japanese syntax, and grammar translated from English grammar. an English language analysis section that analyzes the English sentence input from the input means by referring to the English-Japanese word dictionary and the translation rule dictionary, and an output of the English language analysis section that refers to the translation rule dictionary. an English-Japanese converter that converts the input syntax into the input syntax and provides a translation for each English word; a Japanese generation unit that generates a Japanese sentence; a translation execution unit that translates the original text stored in the original-translation pair management table; and a translation execution unit that associates the original text and the translated text stored in the original-translation pair management table. A translation system comprising a display means for displaying a display, and a Japanese language analysis unit that refers to the English-Japanese word dictionary and analyzes a stored word group of translated words using Japanese grammar to make a determination. A machine translation device featuring: 2. The machine translation device according to claim 1, wherein a Japanese word dictionary with other foreign languages is referred to instead of an English-Japanese word dictionary to translate from the other foreign language into Japanese.