JPH07141375A - Machine translation device - Google Patents

Abstract

(57) [Summary] [Purpose] By setting the most suitable dictionary according to the document to be translated to the most suitable priority and performing the translation process,
(EN) Provided is a translation device which can obtain a translated text of high quality and can reduce the labor of post-editing. [Structure] When the dictionary selection process is activated, a list of usable specialized dictionaries and user dictionaries is displayed (S410), and a flag value indicating priority is set to 0 (S420). When the user sequentially selects the dictionary, the flag value is incremented each time and stored in association with the selected dictionary (S440, S450). The smaller the flag value, the higher the priority. When a dictionary is searched and a translation word with the same part of speech is extracted from a plurality of dictionaries, only the translation word extracted from the dictionary with the highest priority is used.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a machine translation apparatus which uses a plurality of translation dictionaries with priorities set and preferentially uses a dictionary with a high priority.

[0002]

2. Description of the Related Art Conventionally, in this type of machine translation apparatus, a plurality of dictionaries edited according to purposes have been used to perform translation work, and generally, a wide range of words and phrases are generally translated or used. The basic dictionary registered in the translation device and held by the translation device in advance, and the specialized dictionary in which words and translations used only in a specific field are registered and edited for each specialized field, and the user can edit independently. A user dictionary that can be used is used.

When the translated word for a certain phrase is registered in a plurality of dictionaries, the translation process is executed using the translated word registered in the dictionary with the highest priority, and the priority is given to using these dictionaries. The ranking is usually a user dictionary,
The special dictionary and the basic dictionary are fixed in this order.

[0004]

However, when the priority order of the dictionaries is fixed as described above, once the words and phrases are registered in the user dictionary, the same words and phrases are registered in the specialized dictionary and the basic dictionary. However, since the translated word is not used, there is a problem that if the wrong translation is registered in the user dictionary, the correct translated word will not be used thereafter.

Further, even when an external document such as a paper is to be created, it is necessary to perform translation using general words and phrases as much as possible. Since the dictionary priority is fixed, the user dictionary is prioritized. There was also a problem that it was not translated into the desired translation, and post-editing was time-consuming.

The present invention has been made to solve the above-mentioned problems, and when performing a translation work using a machine translation device, the priority order of a plurality of electronic dictionaries is arbitrarily set and translated. The purpose is to preferentially use the dictionary most suitable for a document for translation.

[0007]

The present invention, which has been made to achieve the above object, has an input means for inputting an original sentence in a first language and an entry word in the first language, as shown in FIG. A plurality of dictionaries in which the translated words of the second language corresponding to are registered according to the usage, and the dictionaries with a high priority are preferentially used according to the priorities preset in the dictionaries, and are input by the input means. A plurality of dictionaries prepared in a machine translation device including a translation means for translating an original sentence into a second language and an output means for outputting a translated sentence in a second language obtained by the translation means. A required dictionary is selected from among the selected dictionaries, priority setting means for setting a priority to the selected dictionary, and priority storage means for storing the priority set by the priority setting means are provided. The translation means is And executes a translation process on the basis of the stored priority order storage means.

[0008]

In the machine translation apparatus of the present invention configured as described above, the user selects the required dictionary from the plurality of prepared dictionaries by the priority setting means before performing the translation process, Set priorities for selected dictionaries. Then, the priority order set by the priority order setting means is stored in the priority order storage means.

Then, the translation unit translates the original text of the first language input by the input unit by preferentially using a dictionary with a high priority according to the priority of each dictionary stored in the priority storage unit. The translation processing is performed in two languages, and the translated text in the second language obtained by this translation processing is output by the output means.

[0010]

Embodiments of the present invention will be described below with reference to the drawings. First, FIG. 2 is a block diagram showing a schematic configuration of a Japanese-English machine translation device of an embodiment to which the present invention is applied.

As shown in FIG. 2, the machine translation apparatus of this embodiment is used for a translation process, and an input unit 2 including a keyboard for inputting Japanese sentences to be translated and various operation commands. An external storage device 4 including a hard disk or an optical disk in which various dictionaries are stored, a display unit 6 including a CRT or the like for displaying an input Japanese original sentence and an English translated sentence that is a translation result, and a display unit A display control unit 7 for controlling data displayed on the display unit 6, a printing unit 8 including a laser printer for printing and outputting translated texts of English as a translation result, input from the input unit 2 and the external storage device 4. The control unit 10 executes a translation process based on the data and displays the result on the display unit 6 or outputs the result from the printing unit 8.

Here, in the input unit 2, character keys for inputting characters, input sentence edit keys for activating various processes, translation execution keys, dictionary edit keys, output sentence edit keys, end keys, printing Various function keys such as the execute key, cursor keys used for editing the input and output sentences and operating the screen displayed on the display unit 6, the insert key, the delete key, the enter key, and the next candidate selection key are provided. There is.

In the external storage device 4, as a dictionary used for translation processing, a basic dictionary 12 that the translation device previously holds internally, and a specialized dictionary 1 edited for each specialized field.
4. The user dictionary 16 created by the user is stored. A plurality of dictionaries may exist at the same time for the specialized dictionary 14 for each specialized field and for the user dictionary 16, for example, for each user.

FIG. 3 shows an example of the data structure of the dictionary. FIG. 3A shows the basic dictionary 12, FIG. 3B shows the specialized dictionary 14, and FIG. 3C shows the user dictionary 16. Represent The individual dictionary information includes the entry word 30, the part of speech 32 of the entry word 30, and the translation word 34 corresponding to the entry word 30, respectively.
And the part of speech 36 of the translated word 34.

Next, the control unit 10 includes a CPU 18 and a ROM.
It comprises a well-known microcomputer provided with 20, a RAM 20, an input / output port 24, and a bus line 26 connecting these parts. Then, the control unit 10 translates the Japanese original sentence input from the input unit 2 into an English translated sentence, and at the same time translates and edits the original sentence and translated sentence. A dictionary selection process or the like is performed in which the selected dictionary is given a priority while being selected.

Further, the RAM 22 stores the Japanese original sentence input from the input unit 2 and the English translated sentence translated by the translation editing process, and also stores the priority information set by the dictionary selecting process. The priority information is referred to when searching the dictionary in the translation process.

Next, the translation editing process executed by the control unit 10 will be described with reference to the flow charts shown in FIGS. As shown in FIG. 4, when the translation editing process is executed, first, in S110, the user operates the character keys and the like of the input unit 2 and waits for the input of a Japanese original sentence to be translated. . Here, the input of the original text may be performed by reading an already created original text file. Then, when the Japanese original sentence (hereinafter, input sentence) is input,
In step S120, the input sentence is displayed in a predetermined area on the display unit 6.

In the following S130, it is determined whether or not the input sentence edit key provided in the input section 2 has been operated. If not, the process proceeds to S150. If the input sentence edit key is operated, the process proceeds to S140, and the input sentence edit process is executed. When the input sentence editing process is executed,
The user can directly correct the input sentence displayed in the predetermined area of the display unit 6 by using the cursor key, the delete key, the insert key, etc. provided in the input unit 2.

Next, in S150, it is determined whether or not the translation execution key provided in the input section 2 has been operated. If not, the process proceeds to S190. If the translation execution key has been operated, the process proceeds to S160, a translation process for translating a Japanese input sentence into an English translated sentence is executed, and the subsequent S1
At 70, the output sentence obtained by the translation processing is paired with the input sentence and stored in the RAM 22, and at S180, the input sentence and the output sentence are displayed in predetermined areas provided on the display unit 6, respectively. To do.

The translation process may be executed for all the input sentences at once, or may be specified for each sentence in a range and executed only for the sentences in the specified range.
In S190, it is determined whether or not the dictionary edit key provided in the input unit 2 has been operated. If not, S21
Move to 0. If the dictionary edit key is operated, the process proceeds to S200 and the dictionary edit process is executed. When the dictionary editing process is executed, the user can correct the content of the information registered in the user dictionary 16 or register a new word or phrase.

In S210, it is determined whether or not the output sentence edit key provided in the input section 2 has been operated. If not, the process proceeds to S230. If the output sentence edit key is operated, the process proceeds to S220 to execute the output sentence edit process. When the output sentence editing process is executed, the user outputs the output sentence (English translation) displayed in a predetermined area of the display unit 6 by executing the previous S160 translation process.
Can be directly corrected by using the cursor key, the delete key, the insert key, etc. provided in the input unit 2, as in the input sentence editing process of S140.

In S230, it is determined whether or not the end key provided in the input unit 2 has been operated. If not, the process proceeds to S240. For example, the translated result is output to the printing unit 8 and printed. A key operation such as a print execution key to be detected is detected, and a process corresponding to the operated key is executed.

Then, the above S140, S180, S20
After 0, S220, and S240 are processed, the process returns to S130 and these processes are repeatedly executed. Also, S23
When it is determined that the end key has been operated at 0, this processing ends.

That is, in the translation editing process, after the Japanese original sentence is input, the input sentence editing process, the translation process, the dictionary editing process, and the output sentence editing process are performed according to the operation of various keys provided in the input unit 2. Various processes such as are executed respectively. Then, if the translation execution key is operated and the translation process is executed, but if the obtained output sentence (English translation) is incorrect or unnatural, operate the input sentence edit key. To edit the input sentence by re-editing the original Japanese sentence that is the input sentence, or by operating the dictionary edit key to execute the dictionary edit process to change the contents of the user dictionary 16 and then again. The translation execution key is operated to re-translate, or the output sentence edit key is operated to execute the output sentence edit process to directly correct the English translated sentence which is the output sentence.

If the translation result desired by the user can be obtained by repeating the above processing, the print key is operated to print the translation result and output from the printing unit 8, and then the end key is operated. Then, the process ends. next,
The translation process executed in S160 and translated from the Japanese original sentence (input sentence) to the English translated sentence (output sentence) will be described with reference to the flowchart shown in FIG.

First, in S250, using a predetermined morpheme analysis rule, a dictionary is searched for words forming an input sentence, the words are cut out, and search information such as translated words and parts of speech is added to the morpheme analysis result. Perform analysis processing. Here, the dictionary search processing for searching the words that form the input sentence will be described in detail with reference to the flowchart shown in FIG.

The specialized dictionary 14 and the user dictionary 16 are
A search is performed according to the priority information which is preset and stored in the RAM 22 by a dictionary selection process described later. Further, the priority information stores a flag value indicating the priority set in each dictionary. The dictionary in which the flag value is set to 1 has the highest priority, and the value of the flag becomes large in the following. The lower the priority.

Further, since the basic dictionary 12 includes adverbs and particles necessary for syntax analysis, if the translation process is executed without using the basic dictionary 12, there is a high possibility that the syntax analysis will fail. Therefore, when searching the dictionary, the basic dictionary 12 is always searched.

As shown in FIG. 6, in the dictionary search process, first, in S310, a check value K for checking the flag value given to each dictionary is set to 1 which is an initial value. In subsequent S320, it is determined whether or not there is a dictionary in which a flag value having the same value as the check value K exists, and if such a dictionary exists, the process proceeds to S330 and the same flag value as the check value K exists. The dictionary in which is set is searched, and the translated word corresponding to the headword and its part-of-speech information are read as search information.
Then, in subsequent S330, 1 is added to the value of the present inspection value K, and then the process returns to S320.

On the other hand, if it is determined in the previous S320 that there is no dictionary in which the flag value of the same value as the check value K is set, the process proceeds to S350, the basic dictionary 12 is searched, and the previous S330 is executed. Similarly, the translated word corresponding to the headword and its part-of-speech information are read out as search information.

Then, in the following S360, the search information read in the previous S330 and S350 is examined, and it is judged whether or not there are a plurality of translated words having the same part of speech by being read from a plurality of dictionaries. If there are a plurality of translated parts of speech, the process proceeds to S370, and only the translated word read from the dictionary with the highest priority is retained as the search information, and other translated words are deleted from the search information and the dictionary search processing is performed. finish. Further, in S360, if there are not a plurality of translated words having the same part of speech, the read information is retained as search information and the dictionary search processing for that word is ended.

Then, the morphological analysis processing divides the input sentence into meaningful word strings based on the search information obtained by such dictionary search processing, and adds search information to each word. Returning to FIG. 5, in S260, the result of the morphological analysis process is received, the analysis process for analyzing the Japanese structure is executed using a predetermined analysis grammar, and the subsequent S270.
Then, using a predetermined conversion grammar, the Japanese structure analyzed by the analysis process is converted into an English structure, and the conversion process for replacing each word with English is executed. In subsequent S280, an output sentence composed of an English word string is determined by using a predetermined generation grammar to determine the English word order from the English structure converted by the conversion process, process the tense, the aspect, the person name, etc. Execute the generation process to generate. Then, in S290, by using a predetermined morpheme generation rule, inflectional forms such as plural nouns, past verbs, and past participles are generated for the English word string generated by the generation process, The morpheme generation process for obtaining the output sentence is executed, and the translation process ends.

Next, the dictionary selection processing executed by the control unit 10 will be described with reference to the flowchart shown in FIG. This process allows the user to select whether or not to use the specialized dictionary 14 or the user dictionary 16 prepared in the external storage device 4 and, if so, with what priority. The result is stored in the RAM 20 as priority information.

As shown in FIG. 7, when the dictionary selection process is started by operating the dictionary selection execution key provided in the input unit 2, first, in S410, the usable specialized dictionary 14 and the user are available. The dictionary selection menu in which the list of the dictionaries 16 is listed is displayed on the display unit 6, and in the subsequent S420, a flag value indicating the priority of the dictionary is set to 0.

The flag value 1 has the highest priority, and the higher the value, the lower the priority. In S430, the user operates the next candidate key and the selection key provided in the input unit 2 to determine whether or not the dictionary is selected from the dictionaries displayed in the display unit 6 in S410, If no dictionary is selected, the process proceeds to S460, and
If any dictionary is selected, the process proceeds to S440, the current flag value is incremented by 1, the added value is stored as the flag value of the selected dictionary in S450, and then the process proceeds to S460. move on.

In S460, it is determined whether the user operates the end key provided in the input section 2 to input the selection end command. If the selection end command has not been input, the process returns to S430. On the other hand, if it is determined in S460 that the selection end command has been input, the process proceeds to S470 to end the display of the dictionary list, and then in S480, the set priority information can be used in the translation process. As described above, the priority information in which the information for identifying the selected dictionary is associated with the flag value set in the dictionary is R
The result is stored in the AM 22, and the dictionary selection processing is ended.

That is, the user selects a dictionary from the dictionaries displayed on the display unit 6 in the order in which he / she wants to set it so that it is used preferentially. A flag value indicating the degree is set. For dictionaries not selected in this process,
The flag value is not set and is not used in translation processing. If specialized dictionary 14 and user dictionary 16
If neither of the above is selected, only the basic dictionary 12 will be used in the translation process.

Next, the operation of the machine translation apparatus of this embodiment will be described using a concrete example. First, the dictionary selection process is executed, the specialized dictionary 14 is selected first, and then the user dictionary 16 is selected.
Is selected, the flag value of the specialized dictionary 14 is set to 1 and the flag value of the user dictionary 16 is set to 2. Therefore, the dictionaries used in the translation process are the specialized dictionary 14 and the user dictionary 1.
6, the basic dictionary 12 is given priority in this order.

Next, the user inputs an input sentence from the input section 2 "If this adjustment is not correct, belt loop folding failure and sewing failure will occur." And the input section 2 is continuously provided. When the translation execution key is operated, translation processing is performed. Here, as shown in FIG. 3, "defect folding" and "reject" are registered in the user dictionary 16 and "belt loop" in the specialized dictionary 14. "" Sewing "seaming""
“Defective“ failture ”” and the like are registered, and further, “sew” “sew”, “defective“ rascal ””, and the like are registered in the basic dictionary 12.

In the morphological analysis process, for example, when the word "bad" is cut out, the dictionary search process S310 to S350 results in "bad" failure "" from the specialized dictionary 14 and "from the user dictionary 16". Bad "rej
ect "", "defective" rascal "" is read from the basic dictionary 12 as search information. However, since all of these parts of speech are nouns, the dictionary with the highest priority set in S360 to S370 is read. The translated word issued, that is, the search information other than "defective" failture "" in the specialized dictionary 14 is deleted.

When the word "sew" is cut out, the dictionary search process reads "sew""seaming" from the special dictionary 14 and "sew" from the basic dictionary 12. However, since they have different parts of speech, none of them are retained as search information without being deleted.

In this way, the input sentence is divided into meaningful words, and is passed to the next step, the analysis processing, together with the information such as the part of speech searched for each word by the dictionary search processing. In the analysis processing, by analyzing the input sentence, when information about a plurality of parts of speech is included as search information for one word, such as “sew on”, a more appropriate part of speech is selected. As a result, here, the specialized dictionary 14
"Sew""seaming""is selected, and" sew "" sew "in the basic dictionary 12 is discarded. In the subsequent conversion process, each word is replaced with a translated word, and then the generation process,
By sequentially executing the morpheme generation process, the output sentence "Whe
n this arrangement is bad, defect folding of the b
"elt loop and seaming failture are occurred."

If, in the dictionary selection process, the priorities of the dictionaries are set in the order of the user dictionary, the specialized dictionary, and the basic dictionary, the "bad""reject""in the user dictionary 16 becomes" bad "" failure "in the specialized dictionary 14. "When this arrangement is bad, defe
ct folding of the belt loop and seaming rejectare
occured. "is obtained. In this translation, the defect is" reje
It is not very appropriate to translate it as "ct", so prioritizing specialized dictionaries will give better output.

As described above, according to the machine translation apparatus of this embodiment, the user has the specialized dictionary 14 and the user dictionary 1
Select the dictionary most suitable for the input sentence to be translated from 6,
In addition, priority can be set for the selected dictionary.
Then, according to this priority order, since the translation process is performed by preferentially using the translation word registered in the dictionary with a high priority, it is possible to obtain the translation text of good quality that is suitable for the user's intention.

In other words, the machine translation device is not always used only by a particular user to translate a document of a particular type, but a plurality of users may translate documents of different types and fields. For that purpose, it is necessary to build and prepare specialized dictionaries and user dictionaries for various specialized fields. The present invention is extremely effective in that the machine translation device itself can be used more effectively by utilizing the various prepared dictionaries in the most effective combination.

In this embodiment, the basic dictionary 12
Was configured with the lowest priority, but specialized dictionary 1
4 and the user dictionary 16, it may be configured such that whether or not to use the basic dictionary 12 and with what priority is set.

Further, in this embodiment, the specialized dictionary 14 is used.
The user dictionaries 16 and the user dictionaries 16 are described as being used one by one, but a plurality of specialized dictionaries 14 and user dictionaries 16 may be used at the same time.

[0048]

As described above in detail, in the machine translation apparatus of the present invention, a required dictionary is selected from a plurality of dictionaries registered according to usage, and a priority is set in the selected dictionary. Priority setting means is provided. Then, in accordance with the priority order set by the priority order setting means, the translation process is executed by preferentially using the dictionary having a high priority.

Therefore, according to the machine translation apparatus of the present invention,
The user can set the most suitable dictionary to the most suitable priority according to the document to be translated, and since the translation processing is performed based on the dictionary thus selected and the priority is set. , You can get a good quality translation. In addition, since a translated text of high quality can be obtained, the labor of post-editing the translated text can be reduced.

[Brief description of drawings]

FIG. 1 is a block diagram illustrating a configuration of the present invention.

FIG. 2 is a block diagram showing the overall configuration of a machine translation device of this embodiment.

FIG. 3 is an explanatory diagram showing a configuration of dictionary data stored in an external storage device.

FIG. 4 is a flowchart showing a translation editing process executed by a control unit.

5 is a flowchart showing a translation process of S160 in FIG.

FIG. 6 is a flowchart showing a dictionary search process executed in the morphological analysis process of S250 in FIG.

FIG. 7 is a flowchart showing a dictionary selection process executed by a control unit.

[Explanation of symbols]

2 ... Input unit 4 ... External storage device 6 ... Display unit
7 ... Display control unit 8 ... Printing unit 10 ... Control unit 24 ... Input / output port
26 ... Bus line

Claims (1)

[Claims] 1. An input unit for inputting an original sentence in a first language, a plurality of dictionaries in which translated words in a second language corresponding to headwords in the first language are registered for each purpose, and preset in each dictionary. A translation means for translating the original text input by the input means into a second language by preferentially using a dictionary with a high priority according to the priority order, and a translation means obtained by the translation means. In a machine translation device provided with an output means for outputting a translated text in two languages, selecting a required dictionary from the plurality of prepared dictionaries,
Priority order setting means for setting a priority order in the selected dictionary and priority order storage means for storing the priority order set by the priority order setting means are provided, and the translation means stores the priority order in the priority order storage means. A machine translation device characterized by executing a translation process based on stored priorities.