JPH01100673A - Machine translator - Google Patents

Abstract

PURPOSE:To facilitate the selection by a user by providing a part of speech estimating part for estimating the part of speech of a word and outputting all the estimated part of speech data when there is the data indicating the word which is not stored in a translation dictionary storing part. CONSTITUTION:When an English sentence is inputted from an input device 1, it is stored in an original sentence storing part 2 and a translation dictionary retrieving part 4 refers to the translation dictionary storing part 3 according to an English/Japanese translation instruction to retrieve the dictionary. When there is the data which is not stored in the storing part 3, the part of speech estimating part 7 selects one part of speech to instruct the execution of analyzing a syntax to a parsing syntatic analyzing part 9. At this time, the analyzing part 9 refers to a syntax dictionary storing part to decide whether it is established as the sentence or not. When the syntax is not established, the part of speech is changed to other part of speech to recheck whether the syntax is established or not. When the syntax is established, the part of speech is registered to a part of speech virtually registering part 15. When the user selects a specific part of speech, the data indicating the part of speech selected by the storing part 3 is stored to reduce the burden of the user.

Description

DETAILED DESCRIPTION OF THE INVENTION Field of the Invention The present invention relates to a machine translator used to translate from one language to another, such as from English to Japanese.

BACKGROUND OF THE INVENTION In recent years, machine translators have become widely used as translation support devices.

Generally, not all words are registered in the dictionary of a machine translator, and special terms are often unknown. Therefore, the user must register words that are not registered in such dictionaries.

A conventional example of word registration is a method in which the spelling of the original language, grammatical information, translated words, etc. are registered interactively between a user and a translator.

Problems that the invention aims to solve However, inputting grammatical information requires a very high level of grammatical knowledge, and some machine translators have an intermediate language that is different from the natural language. Since it is necessary to register grammatical information specific to the intermediate language, it is difficult for users to understand, and this has been a major obstacle when registering words.

Means for Solving the Problems In order to solve the above problems, the present invention provides that when there is data representing a word that is not stored in the translation dictionary storage unit in the data representing the original text, the part of speech of this word is The system is equipped with a part-of-speech estimator that estimates the part-of-speech and outputs all the estimated part-of-speech data.

Effect of the Invention Since the present invention is comprised of the above-described means, when there is data representing a word that is not stored in the translation dictionary storage section, the part-of-speech estimating section outputs all grammatically valid parts of speech.

Embodiment An English-Japanese machine translator which is an embodiment of the present invention will be described with reference to the drawings.

FIG. 1 is a block diagram of an English-Japanese machine translator in one embodiment of the present invention.

In FIG. 1, 1 is a means for inputting data representing an English sentence, a means for instructing to register the word data in a translation dictionary storage unit (to be described later) when the sentence contains an unregistered word, and a means for translating. This is an input device that also has means for instructing the start of execution. Reference numeral 2 denotes an original text storage unit that stores data representing an English text input from an input device. 3 is a translation dictionary storage unit that stores grammatical information such as the notation 9 parts of speech of English and Japanese words and phrases in correspondence; 4 refers to the translation dictionary storage unit 3 for words and phrases in the input English text; This is a translation dictionary search unit that retrieves information necessary for translation. Reference numeral 5 denotes a word registration unit which, when an unregistered word exists in the input sentence, interacts with the user to import necessary information and update the translation dictionary storage unit 3. Reference numeral 6 denotes a translation information storage unit that stores information taken in by the translation dictionary search unit 4. Reference numeral 7 denotes a part-of-speech estimating unit that estimates grammatical information such as the part of speech while interacting with a syntactic analysis unit, which will be described later, when registering a word. Reference numeral 8 denotes a syntax dictionary storage unit that stores English syntax, and 9 a syntactic analysis unit that performs analysis by referring to the syntax dictionary storage unit 8 and the translation information storage unit 6. 10 is a syntactic information storage unit that stores the results of syntactic analysis;
Reference numeral 11 denotes a translation generation dictionary storage unit that stores Japanese grammatical information necessary to generate a translation from the result of syntactic analysis.

12 is a translation generation unit that generates a translation by referring to the translation generation dictionary storage unit 11 and the syntax information storage unit 10; 13 is an output storage unit that stores the generated translation; and 14 is a translation stored in the output storage unit. This is a display device that displays instructions to the user. 15
is a temporary part-of-speech registration unit that temporarily stores the part-of-speech data output by the part-of-speech estimation unit 7.

The operation of the English-Japanese machine translator configured as described above will be explained according to the flowchart shown in FIG.

First, in step A, when an English sentence is input from the input device 1, the English sentence is stored in the original sentence storage section 2.

Furthermore, when a key for executing English-Japanese translation is input from the input device 1, the process moves to step B. In step B, the translation dictionary search unit 4 searches the dictionary for words and phrases in the input sentence while referring to the translation dictionary storage unit 3, and retrieves grammatical information such as Japanese expressions and parts of speech for each word and phrase. If there is a word in the input sentence that is not registered in the dictionary, the fact that the word is an unregistered word is stored in the translation information storage section 6, and the process moves to step C.

In step C, it is determined by referring to the translation information storage section 6 whether or not there are any unregistered words in the input sentence. If an unregistered word exists in the input sentence, the process moves to step D, and if no unregistered word exists in the input sentence, the process moves to step G.

Hereinafter, an example will be described in which the sentence "rl 5tudy achaeology" is input. r achaeolo
If gy J is an unregistered word, follow steps A-C above.
Step D is then reached.

In steps DF, words are registered by interacting with the user. That is, in step D, the word registration unit 5 says, ``There is an unregistered word.The unregistered word is 'achae.

1ogy". Would you like to correct the spelling of this word?

A message such as (Y/N) J is displayed on the display device 14. Thereafter, in step E, if the unregistered word is not a spelling mistake by the user, that is, if the user has not corrected the spelling, control is transferred to the part of speech estimating section 7, and the process proceeds to step F. If the user corrects the spelling,
Go back to step B and repeat the above process again with the corrected spelling.

In step F, the part-of-speech estimating unit 7 first selects one part of speech from among a plurality of parts of speech for unregistered words, and instructs the syntactic analysis unit 9 to perform syntactic analysis.

At this time, the syntactic analysis section 9 refers to the syntax dictionary storage section 8 and determines in step G whether "l 5 study achaeology" is valid as a sentence. If the syntax does not hold, change the part of speech to another part of speech and check again to see if the sentence holds. If the syntax is established, step H
The part of speech is registered in the part of speech provisional registration section 15. After this registration, check whether setting of all parts of speech is completed in step I, and step F until setting of all parts of speech is completed.
Repeat the following process. In this way, we perform syntactic analysis for all parts of speech to find out which part of speech the syntax is successful in,
All parts of speech that form the syntax are registered in the temporary part of speech registration unit 15. The display device 14 displays “The part of speech of this word is 1, noun, 2.
．． It is presumed to be an adverb. Please choose the appropriate part of speech.

” is displayed to instruct the user to select the part of speech. When the user selects a specific part of speech, data representing the selected part of speech is stored in the translation dictionary storage section 3 in step J. In addition, the display device 14 displays "Please input the translated word." and data representing the translated word input by the user is stored in the translation dictionary storage section 3.

In step, the syntax analysis section 9 performs syntax analysis while referring to the syntax dictionary storage section 8, and writes the result to the syntax information storage section 10. Next, the translation generation section 12 generates a translation while referring to the syntax information storage section 10 and the translation generation dictionary storage section 11, and writes it into the output storage section 13. Furthermore, in step L, the display device 14
The contents of the output storage section 13 are displayed and the process ends.

Effects of the Invention According to the present invention, when there is data representing a word that is not stored in the translation dictionary storage unit in the data representing the original text,
There is a part-of-speech estimation unit that estimates the part-of-speech of this word and outputs all the estimated part-of-speech data, so even if there is an unregistered word in the input sentence, the part-of-speech estimation result is displayed, so the user can check the estimated part-of-speech. By selecting one of these, the part of speech of an unregistered word can be registered, which reduces the burden on the user. Furthermore, if part-of-speech estimation is performed by collecting example sentences in which an unregistered word is used in the same manner using the effect, the user does not need to have much knowledge of grammar.

[Brief explanation of the drawing]

FIG. 1 is a block diagram of a machine translator according to an embodiment of the present invention, and FIG. 2 is a flowchart showing the same word registration operation. 1... Input device 2... Original text storage unit 3...
... Translation dictionary storage section 4 ... Translation dictionary search section 5 ... Word registration section 6 ... Translation information storage section 7 ... Part of speech estimation section 8 ... Syntax dictionary storage section 9 ...Syntax analysis unit 10...Syntax information storage unit 11...Dictionary storage unit for translation generation

Claims (1)

[Claims] an original text storage unit that stores data representing the original text, a translation dictionary storage unit that stores data representing parts of speech and translated words of words in the original text, and a syntactic analysis unit that analyzes the syntax of the original text; If there is data representing a word that is not stored in the translation dictionary storage unit, a part of speech estimation unit that estimates the part of speech of this word and outputs all data representing candidate parts of speech; and output data of the part of speech estimation unit. 1. A machine translation machine comprising: a display means for displaying the part of speech; and an input means for instructing one of the output data of the part-of-speech estimator displayed on the display means.