JPH06162079A - Parallel translation retrieving device - Google Patents

Abstract

PURPOSE:To realize a good retrieval efficiency without depending on an operator's input operation by extracting an independent word which can become a retrieval keyword from in an inputted sentence, and converting it to second language by using a word parallel translation data store means. CONSTITUTION:A user inputs a sentence of first language from an input part 101, and the inputted sentence of first language is held in an input holding part 102. Also, the sentence of first language is divided into word units, while referring to a word dictionary 104, and in a word train of the sentence of divided first language, that which has semantics as a word, for instance, an independent word is selected and held in a word train holding part 105. Subsequently, a retrieval keyword shown by second language for retrieving a parallel translation data base is derived. The retrieval keyword is derived by retrieving a translated word of second language corresponding to a word of first language held in a word train combination holding part 107 from in a translated word dictionary 109. Next, by using this retrieval keyword, a parallel translation data base 110 is retrieved, and all parallel translation sentences containing this retrieval keyword are extracted from the parallel translation data base.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a structure of a parallel translation search device for searching a parallel translation database for parallel translations of a sentence written in a certain language and displaying the translations.

[0002]

2. Description of the Related Art For example, there is a parallel translation search device as a kind of translator that translates sentences in different languages, such as converting Japanese sentences into foreign language sentences. This parallel translation retrieval device holds various sentences displayed in a language to be converted as a database in advance, and retrieves and displays a desired parallel translation sentence from a database according to a keyword input by a user.

FIG. 7 is a block diagram showing the structure of a conventional parallel translation search device. As shown in FIG. 7, the parallel translation search device includes a keyword input unit 301 and a keyword holding unit 3.
02, a parallel translation search unit 303, and a parallel translation database 304
And a display unit 305. Keyword input part 3
01 is for the user to select a sentence written in the first language and input the keyword converted into the second language. The keyword holding unit 302 includes the keyword input unit 3
The second language keyword input from 01 is held.
The parallel translation search unit 303 searches the parallel translation database 304 for a parallel translation sentence including a keyword. The bilingual database 304 holds various bilingual example sentences written in the second language. The display unit 305 displays the bilingual sentence searched by the bilingual search unit 303.

Next, the operation of this conventional parallel translation search device will be described. Here, for convenience of explanation, the case where the first language is Japanese and the second language is English, and an example of searching an English parallel sentence for the Japanese sentence "Today is very hot." First, the operator selects a word that is a keyword for a search from among Japanese sentences and further converts it into English. In this case, select “Today”, “Very”, and “Hot” and select “Today”, “Very”, and “ho” respectively.
t ".

After this process is completed, the operator inputs the keywords converted into English individually or in combination with each other from the keyword input unit 301. After the input keyword is held in the keyword holding unit 302,
It is given to the parallel translation search unit 303. The parallel translation search unit 303
The bilingual database 304 is searched, and the English bilingual sentence containing the given keyword is extracted from the bilingual database and output to the display unit 305. The output unit 305 displays the extracted bilingual sentence on the screen. The operator searches for a desired sentence from the displayed bilingual sentences and ends the process when the corresponding sentence is found.

If a parallel translation including a keyword does not exist in the parallel translation database 304 or if an appropriate parallel translation corresponding to a Japanese sentence does not exist in the search result, another English keyword is input again and the search is performed again. To do. If the number of parallel translations extracted by the search is large, a combination with another keyword is input, and the search operation is repeated so as to reduce the number of parallel translations of the corresponding search result. By repeating such an operation, an English bilingual sentence, for example, "It i
Search for "every hot today".

[0007]

However, in such a conventional parallel translation search apparatus, it is premised that the operator inputs a word in the second language which is a search keyword when searching for a parallel translation. For this reason, first, the operator needs to check the word in the second language corresponding to the first language. Further, the operator needs to determine in what combination the keywords are to be input, and the speed of the search processing greatly varies depending on the input method. Furthermore, if the keyword of the second language is inappropriate, it may not be possible to search for an appropriate bilingual sentence, and it is necessary to prepare and re-input an appropriate keyword.

As described above, the conventional bilingual retrieval apparatus has a problem that the operator's burden in the retrieval process is heavy and the speed and accuracy of the retrieval process are affected by the skill of the operator's instruction. Therefore, the present invention has been made in order to solve the above problems, and an object of the present invention is to provide a parallel translation search device that does not depend on the input operation of the operator and has good search efficiency.

[0009]

A parallel translation search device according to the present invention is an input means for inputting a character string written in a first language, and an independent word from the input character string of the first language. Independent word extracting means for extracting, word parallel translation data storage means for storing parallel translations of the independent word written in the first language,
Translated word extraction means for extracting a translated word written in the second language corresponding to the independent word of the first language from the word parallel translation data storage means, and a translated sentence storing various sentence data written in the second language. A storage unit, a parallel translation sentence search unit for extracting from the parallel translation sentence storage unit a parallel translation sentence written in a second language that includes the translated word extracted by the translation word extraction unit, and a display unit for displaying the searched parallel translation sentence. ing.

Further, the parallel translation search device of the present invention further comprises:
An independent word combination generation means for obtaining and holding all combinations of independent words extracted by the independent word extraction means is provided, and the parallel text search means corresponds to the combination of independent words generated by the independent word combination generation means. All bilingual sentences including translated words are extracted from the bilingual sentence storage means.

[0011]

In the parallel translation search device of the present invention, the operator
The sentence for which the translation is requested can be input in the first language.
Then, the independent word extracting unit extracts an independent word that can be a search keyword from the input sentence and converts it into the second language by using the word parallel translation data storage unit. By this processing, the search keyword in the second language is automatically generated. Further, the parallel translation text search means searches the stored data of the parallel translation text storage means using the search keyword in the second language to extract the parallel translation text including the search keyword.

Further, the independent word combination generating means finds all combinations of the search keywords in the second language. Then, the parallel translation sentence search means performs the search operation based on the combination of all the search keywords.

[0013]

Embodiments of the present invention will be described below in detail with reference to the drawings. FIG. 1 is a block diagram showing the configuration of a parallel translation search device according to an embodiment of the present invention. Referring to FIG. 1, this parallel translation search device is realized by, for example, a word processor, a personal computer, or the like, and is configured by a reading device from a floppy disk or a hard disk, an input unit 101 configured by a keyboard, and a RAM. Input sentence holding unit 102, word string holding unit 105,
A combination holding unit 107, a word dictionary 104 including a ROM and a RAM, a translated word dictionary 109, a parallel translation database 110, a display unit 111 including a CRT and a printer, and a built-in parallel translation search apparatus. A word division unit 103, a combination generation unit 106, and a parallel translation search unit 108 that operate under the control and judgment of a microcomputer are provided.

The input unit 101 is for the user to input a sentence in the first language. The input sentence holding unit 102
The first language sentence input to the input unit 101 is stored and held. The word dividing unit 103 divides the sentence of the first language held in the input sentence holding unit 102 into word units. Here, the word refers to a group of characters that can be classified by each part of speech.

The word dictionary 104 holds word information of the first language used when the word dividing unit 103 divides the sentence held in the input holding unit 102 into word units. The word string holding unit 105 holds the word string of the sentence of the first language divided by the word dividing unit 103 into words. The combination generation unit 106 generates a combination of words in the first language held in the word string holding unit 105.

The combination holding unit 107 holds the word combination of the first language generated by the combination generation unit 106. The combination number holding unit 112 holds the number of word combinations of the first language generated by the combination generation unit 106. The parallel translation search unit 108 obtains a translated word in the second language of the set of words in the first language held in the combination holding unit 107, and searches a parallel translated database in the second language including the translated word in the parallel translation database.

The search count holding unit 113 is a parallel translation search unit 10.
8 holds the number of combinations held in the combination holding unit 107 retrieved. The translated word dictionary 109 holds translated word information of the second language for the word set of the first language held in the combination holding unit 107. Bilingual database 11
0 holds various parallel translations written in the second language.

The display unit 111 displays the bilingual sentence in the second language retrieved by the bilingual retrieval unit 108. Next, the operation of this parallel translation search device will be described. FIG. 2 is a flowchart showing the operation procedure of this parallel translation search device, and FIG. 3 is a flowchart showing the detailed operation procedure of the parallel translation search step and the parallel translation display step of FIG. Further, in the following description, in order to make it easier to understand the difference from the conventional parallel translation search device, the same Japanese example sentence “Today is very hot.”
An example will be described in which the English parallel translation is searched.

First, the first language input step (step 20)
In 1), the user inputs a sentence in the first language from the input unit 101. Next, input sentence holding step (step 202)
Then, the input sentence storage unit 102 stores the input first language sentence.
Hold on. Further, word division step (step 203)
Then, the sentence in the first language is divided into word units for recognition with reference to the word dictionary 104. For example, as shown in FIG. 4, in the word dictionary 104, "today" is a noun, "ha" is a particle, "very" is an adverb, and "hot" is an adjective, and the part-of-speech information of each word, The particle information stores connection information (not shown) between words that connects to a noun. Then, the word dividing unit 103 uses the word dictionary 104.
With reference to, the input sentence "Today is very hot." Is recognized by being divided into four words "Today", "Ha", "Very", and "Hot".

In the word string holding step (step 204),
Of the divided word strings of the sentence in the first language, one having meaning as a word, for example, an independent word is selected and the word string holding unit 1 is selected.
Hold at 05. In the above example of the input sentence, the word strings “today”, “very” and “hot” are retained. The particle "ha" is
Since it is an auxiliary word that has no meaning by itself, it is not stored as a word string.

In the combination generation step (step 205), the word combination held in the word string holding unit 105 is generated. In the above example, "Today" and "Very"
Seven combinations of “hot,” “today, very,” “very, hot,” “today, hot,” “today, very hot,” are generated. The seven sets of word strings are combined and stored in the combination holding unit 1.
The number of combinations is held at 07, and the number of combinations, 7, is held at the number-of-combinations holding unit 112 (step 206).

Next, the process moves to the parallel translation search stage (step 207). Referring to FIG. 3, first, a search keyword expressed in the second language for searching the parallel translation database is obtained. The search keyword is obtained by searching the translated word dictionary 109 for a translated word in the second language corresponding to the word in the first language held in the word string combination holding unit 107. FIG. 5 shows the contents of the translated word dictionary 109 including the searched translated words. Then, the translated words shown in FIG. 5 are used as search keywords.

When the search keyword is obtained, the parallel translation database 110 is searched using this search keyword,
All bilingual sentences including this search keyword are extracted from the bilingual database. Various methods can be applied to the search method at this processing stage in connection with the construction of the parallel translation database. As an example, a case where the parallel translation database has a data area for storing various parallel translation example sentences and an index data area for storing the words included in the parallel translation sentences and the addresses of the parallel translation sentences will be described. First, the index data of the parallel translation database is searched for the search keyword “today”. When a word that matches the keyword is detected, the bilingual sentence is extracted by referring to the address data of the bilingual sentence stored together with the word. Further, the search operation of the next search keyword "very" is started. When the search is completed for all the search keywords, the extracted bilingual sentence is held in association with the search keyword and the word of the input sentence. FIG. 6 conceptually shows the retained state of the retrieved bilingual sentences.

Next, the parallel translation display stage is entered (step 20).
9). Two display methods can be selected. One method is to sequentially display all extracted bilingual sentences. In this case, of the parallel sentences shown in FIG.
All parallel translations extracted for a single keyword,
In this example, the words “today”, “very”, “hot” (search keywords “today” “very” “hot” “wa”
All parallel translations including "rm") are displayed on the screen of the display unit 111. If the correct English bilingual sentence corresponding to the Japanese input sentence is registered in the bilingual translation database 110, the correct bilingual sentence is displayed in the displayed bilingual sentence. Therefore, the user can find the parallel translation in the displayed document.

The second display method is a method of sequentially displaying each combination of words, that is, each type of combination of search keywords. For example, first, the search results of a combination with a large number of search keywords are displayed. In the example shown in FIG. 6, the parallel translation of combination No. 7 consisting of three words (today is very hot) is displayed. In this example, the corresponding bilingual sentence does not exist in the bilingual database. In this case, the search result for the next combination of words is displayed. Then, such a display operation is repeated until the operator finds a desired bilingual sentence. At the same time, the number of combinations displayed is counted, and when the number of combinations held in the combination number holding unit 112 matches, the process ends.

In the above embodiment, an example in which all parallel sentences including individual search keywords are searched and then displayed in consideration of the combination of search keywords has been described. However, for each individual combination of search keywords You may make it search and display a result for every individual search operation. For example, with respect to the above example sentence, first, the parallel translation database is searched for the search keyword “today” of the word “today”. Then, the search result is displayed on the display unit 111. At the same time, the number of searches is set to 1 and the number of searches holding unit 11
Hold at 3. Then, looking at the search result, when the operator gives an instruction to continue the search, the search processing is continued for the search keyword of the set of the next word combination. When the display of the search result for each combination is completed, the number of searches and the number of combinations are compared. If both match, all parallel translation search processing is terminated.

In the above embodiment, the parallel translation search between Japanese and English has been described, but the present invention is not limited to these languages. The same applies to other languages.

[0028]

As described above, the parallel translation search device of the present invention is
Since the bilingual translation of the word selected from the sentence input in the first language is used as the search keyword and the bilingual translation database is searched for all the combinations to extract the bilingual translation sentence, the operator can use the first language in the first language. By displaying the bilingual sentence only by the operation of inputting the sentence, the retrieval efficiency of the bilingual sentence is improved.

[Brief description of drawings]

FIG. 1 is a block diagram showing the configuration of a parallel translation search device according to an embodiment of the present invention.

FIG. 2 is a flowchart showing an operation procedure of the parallel translation search device shown in FIG.

FIG. 3 is a flowchart showing an operation procedure of a parallel translation search step and a parallel translation display step in the operation procedure shown in FIG.

FIG. 4 is a word dictionary data diagram showing an example of a word dictionary of the parallel translation search device shown in FIG.

5 is a translation word dictionary data diagram showing an example of a translation word dictionary of the parallel translation search device shown in FIG.

6 is a search result display diagram conceptually showing an example of parallel translation search results of the parallel translation search device shown in FIG.

FIG. 7 is a block diagram showing a configuration of a conventional parallel translation search device.

[Explanation of symbols]

 101 input unit 102 input sentence holding unit 103 word division unit 104 word dictionary 105 word string holding unit 106 combination generation unit 107 combination holding unit 108 bilingual search unit 109 translated word dictionary 110 bilingual database 111 display unit

Claims (2)

[Claims] 1. A parallel translation search device for searching and displaying a parallel translation of a second language corresponding to a character string written in the first language, wherein a text string written in the first language is input. Input means, independent word extracting means for extracting an independent word from the input character string of the first language, and word parallel translation data storing parallel translation data of the second language for various words of the first language Storing means, translated word extracting means for extracting a translated word in a second language corresponding to the extracted independent word from the word bilingual data storage means, and bilingual translation storing various sentence data written in the second language. A parallel translation search including a sentence storage unit, a parallel translation sentence search unit for extracting a parallel translation sentence including the translated word extracted by the translation word extraction unit from the parallel translation sentence storage unit, and a display unit for displaying the searched parallel translation sentence. apparatus. 2. The parallel translation search device further comprises independent word combination generation means for obtaining and holding all combinations of the independent words extracted by the independent word extraction means, and the parallel text search means for the independent text search means. 2. The parallel translation search device according to claim 1, wherein all parallel translation sentences including translated words corresponding to the combination of independent words generated by the word combination generating means are extracted from the parallel translation sentence storage means.