JP2007026256A - Word-to-word correspondence determining program and method for parallel translation example dictionary of first language and second language - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a word-to-word correspondence determining program for a parallel translation example dictionary of a first language and a second language that can make words, which are not made to correspond to each other even when a parallel translation word dictionary between the first language and second language is used, correspond to each other between pluralities of words constituting a sentence of the first language and a translation of the second language which are previously made to correspond to each other in the parallel translation example dictionary. <P>SOLUTION: The word-to-word correspondence determining program for the parallel translation example dictionary of the first language and second language according to the present invention performs processing for determining correspondence between words of a pair of a sentence and a translation, specified in the parallel translation example dictionary of the first language and second language in which correspondence is set between sentences of the first language and translations of the second language, by using a parallel translation word dictionary in which words of different languages are related to correspond to each other. <P>COPYRIGHT: (C)2007,JPO&INPIT

Description

  The present invention relates to an interword correspondence determination program and method for a bilingual example dictionary between a first language (for example, Japanese) and a second language (for example, Chinese), for example, a bilingual word dictionary for a first language and a second language. And a bilingual word dictionary between a first language and a third language other than the second language (for example, English), and a program for determining the correspondence between words in the bilingual example dictionary of the first language and the second language And methods.

  2. Description of the Related Art Conventionally, there is a bilingual example dictionary for a first language and a second language in which a correspondence relationship is established between a sentence in a first language and a translated sentence in a second language corresponding to the sentence. The conventional word correspondence recognition technology between the sentence and the translated sentence is based on the correspondence between the sentence and the translated sentence. For example, in order to recognize a corresponding word from a bilingual example sentence in a second language for a sentence in the first language, a bilingual translation between the first language and the second language for each word resulting from morphological analysis of each bilingual example sentence If a word dictionary is applied and the parallel translation word candidate of the word A constituting the sentence in the first language matches the word B constituting the sentence in the second language, it is determined that there is a possibility that A and B correspond to each other. To do. However, the coverage of bilingual word candidate generation by this method depends on the development status of the bilingual word dictionary between the first language and the second language. In an environment where the bilingual word dictionary is not sufficiently developed, the recognition error of the bilingual word is not possible. Is known to occur at a high frequency. For example, when applying a Japanese-English word dictionary in which a vocabulary in the order of one million words is applied, recognition failure is relatively small, whereas when applying a daytime word dictionary in the order of tens of thousands of words, the vocabulary is insufficient. There is a very high frequency of recognition failure. Although word correspondence information is the most useful information among translation support information, it can be said that it provides sufficient information as a translation support system because the recognition accuracy of the corresponding words is poor in languages other than English. First, it has been a problem to improve the word recognition accuracy.

Although there is no direct relationship with the technique of providing a correspondence between words constituting each example (example sentence) of the bilingual example dictionary, Patent Document 1 discloses a bilingual dictionary that provides a correspondence between words in the first language and the second language. And a bilingual dictionary for creating correspondence between the words in the second language and the third language, and creating a bilingual dictionary for the first language and the third language.
Japanese Patent Laid-Open No. 2-301868 “Dictionary Creation / Maintenance Method for Natural Language Processing”

  An object of the present invention is to use a bilingual word dictionary between a first language and a second language between a plurality of words constituting a first language sentence and a second language translation sentence that are associated in advance as a bilingual example dictionary. It is another object of the present invention to provide a program and a method for determining the correspondence between words in the bilingual example dictionary of the first language and the second language, which makes it possible to make correspondence between words that are not matched.

  According to the first aspect of the present invention, the word-to-word correspondence determination program of the first language and second language parallel translation example dictionary sets the correspondence between the first language sentence and the second language translation sentence. In the parallel translation example dictionary between the first language and the second language, a process for determining the correspondence between the words constituting the sentence in the first language and the translation constituting the translation in the second language is made to correspond between the words in different languages. In a program to be executed by a computer using a bilingual word dictionary, using the bilingual word dictionary from the first language to the third language, extracting a word in the third language corresponding to a word in the first language 1, extracting a first language word corresponding to the extracted third language word using a bilingual word dictionary from the third language to the first language, and from the first language, the step Using a bilingual word dictionary for the second language Extracting a second language word corresponding to the first language word in step 1 or the first language word extracted in step 2 as a translation word candidate; and Determining that the candidate word is a translated word corresponding to the word in the first language (in step 1) when the word is included in the word constituting the translated sentence in the language. A program for determining correspondence between words in a bilingual example dictionary between a first language and a second language as a feature.

  Here, the above-described series of extraction processes makes it possible to expand the range of words in the first language that should be candidates for correspondence with the words in the second language. Even if a bilingual word dictionary from a language to a second language is used, the corresponding second language word does not exist, or the corresponding second language word does not exist in the second language translation. Even in such a case, for a certain first language word included in the expanded first language word range, a corresponding second language word may be present in the translation of the second language. Therefore, it is possible to expand and add correspondences between a plurality of words that respectively constitute the sentence in the first language and the translated sentence in the second language.

  According to the second aspect of the present invention, the inter-word correspondence determination program for the first language and the second language parallel translation example dictionary sets the correspondence between the second language sentence and the first language translation sentence. In the parallel translation example dictionary between the first language and the second language, a process for determining the correspondence between the words constituting the sentence in the second language and the translation constituting the translation in the first language is associated with the words in the different languages. In a program to be executed by a computer using a bilingual word dictionary, a word in the first language corresponding to a word in the second language is converted into a first translated word using a bilingual word dictionary from the second language to the first language. Extracting as a candidate; extracting a word in a third language corresponding to a first translated word candidate of the extracted first language word using the bilingual word dictionary from the first language to the third language; , From the third language to the first language Extracting the first language word corresponding to the extracted third language word as a second translation word candidate using the bilingual word dictionary, and the first translation word candidate or the second translation word candidate is the first translation word candidate. If it is included in a word constituting a translation of one language, the first translation candidate or the second translation candidate is determined as a translation corresponding to the second language word (in the first translation candidate extraction step). A program for determining the correspondence between words in a bilingual example dictionary between a first language and a second language.

  Here, by the series of extraction processes described above, it is possible to expand the range of words in the first language that should be candidates for correspondence with the words in the second language. For the initial first language word (candidate) range corresponding to the second language word, the corresponding first language word does not exist in the first language translation, A certain first language word included in the expanded first language word range corresponding to the second language word may be present in the translation of the first language. Therefore, it is possible to expand and attach the correspondence between a plurality of words that respectively constitute the second language sentence and the first language translation sentence.

  According to the third aspect of the present invention, the inter-word correspondence determination program of the first language and second language parallel translation example dictionary sets the correspondence between the first language sentence and the second language translation sentence. In the parallel translation example dictionary between the first language and the second language, a process for determining the correspondence between the words constituting the sentence in the first language and the translation constituting the translation in the second language is made to correspond between the words in different languages. In a program to be executed by a computer using a bilingual word dictionary, step 1 for extracting a semantic symbol corresponding to the word in the first language, and a step for extracting a word in the first language corresponding to the extracted semantic symbol 2 and a word corresponding to the first language word in step 1 or the first language word extracted in step 2 using a parallel translation word dictionary from the first language to the second language. Translate words in two languages And when the translation candidate is included in a word constituting the translation in the second language, the translation candidate is determined as a translation corresponding to the word in the first language in step 1 A program for determining the correspondence between words in a bilingual example dictionary between a first language and a second language.

  According to the fourth aspect of the present invention, the inter-word correspondence determination program for the first language and second language parallel translation example dictionary sets the correspondence between the second language sentence and the first language translation sentence. In the parallel translation example dictionary between the first language and the second language, a process for determining the correspondence between the words constituting the sentence in the second language and the translation constituting the translation in the first language is associated with the words in the different languages. In a program to be executed by a computer using a bilingual word dictionary, a word in the first language corresponding to a word in the second language is first used by using a bilingual word dictionary from the second language to the first language. Extracting as a translation candidate, extracting a semantic symbol of the extracted first language word, extracting a first language word corresponding to the extracted semantic symbol as a second translation candidate, First translation candidate When the second translated word candidate is included in a word constituting the translated sentence of the first language, the first translated word candidate or the second translated word candidate is designated as the second translated word candidate (first translated word candidate extracting step). A program for determining the correspondence between words in a bilingual example dictionary between a first language and a second language, wherein the computer executes the step of determining as a translation corresponding to a word in a language.

  According to the present invention, among the correspondence relationships between a plurality of words constituting a sentence in the first language and a translated sentence in the second language, the bilingual word dictionary of the first language and the second language can be used initially. It is possible to attach a correspondence relationship to those for which no correspondence relationship has been attached.

Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings. Here, a case where the first language is Japanese, the second language is Chinese, and the third language is English will be described.
FIG. 1A is a block diagram illustrating a configuration of a word correspondence determination processing unit of a bilingual example dictionary between Japanese and Chinese according to an embodiment of the present invention.

  As shown in FIG. 1A, the inter-word correspondence determination processing unit 10 of the parallel translation example dictionary 10 sets the correspondence between the Japanese sentence and the Chinese translation sentence, and the bilingual example dictionary between Japanese and Chinese. Process to determine the correspondence between words in a specified sentence / translation pair, a bilingual word dictionary that provides correspondence between words from Japanese to Chinese, and correspondence between words from Chinese to Japanese This is a processing unit that selectively executes a bilingual word dictionary that establishes a relationship, a bilingual word dictionary that establishes a correspondence relationship between words from Japanese to English, and the like.

  In FIG. 1A, the control part 18 is the 1st expansion process execution part 11, the 2nd expansion process execution part 12, the 3rd expansion process execution part 13, the 4th expansion process execution part 14, the 5th expansion process execution part 15, or In the sixth expansion processing execution unit 16, an activation instruction is issued to two or more processing units designated in advance as to be executed.

  The 1st expansion process execution part 11, the 2nd expansion process execution part 12, the 3rd expansion process execution part 13, the 4th expansion process execution part 14, the 5th expansion process execution part 15, or the 6th expansion process execution part 16 Then, by giving a word as a search key to the search unit 17 and executing a necessary search, a process of extending a word (candidate) range of another language corresponding to the word of the designated language is executed. That is, the search unit (extraction unit) 17 extracts (searches) a word in another language or a semantic symbol corresponding to the word in the specified language, using the parallel translation word dictionary specified by each expansion processing execution unit. Or a word in a specified language corresponding to a specified semantic symbol is extracted (searched). After the expansion (enlargement) process is executed, the control unit 18 is notified that the expansion process has been completed, and the expanded word (candidate) range is passed to the hit determination unit 19.

  The hit determination unit 19 includes a word included in an expanded word (candidate) range of another language corresponding to a word in the specified language, which constitutes an example of the other language (pre-specified) Whether the word included in the expanded range hits any of the words constituting the example. After the hit determination process is completed, the hit determination unit 19 notifies the control unit 17 that the hit determination process is completed, and the determination result (if there is a hit word as a corresponding word) It passes to the control unit 18. The number of hit words is usually “1”, but there may be a plurality of words.

  For example, the control unit 18 executes a corresponding word provisional determination process corresponding to the corresponding word candidate expansion process performed immediately before, that is, the first corresponding word temporary determination unit 21 to the sixth corresponding word temporary determination unit 26. An activation instruction is issued to any of the above.

  And by the corresponding unit, based on the determination result by the hit determination unit 19, there is a correspondence between a plurality of words constituting the Japanese example and a plurality of words constituting the Chinese example. A mark writing process to a correspondence matrix (corresponding information) that holds information indicating whether or not is performed is executed.

  The uniqueness determination unit 27 controls writing to the correspondence matrix so that words constituting the Chinese example uniquely correspond to individual words constituting the Japanese example. In this case, as shown in FIG. 1A, after the writing to the correspondence matrix is executed by the first corresponding word tentative determination unit 21 to the sixth corresponding word tentative determination unit 26, the uniqueness determination unit 27 performs the above unique processing. The correspondence relationship may be regenerated so as to satisfy the characteristics, and as shown in FIG. 1B, the first corresponding word tentative determination unit 21 to the sixth corresponding word tentative determination unit 26 perform the writing to the corresponding matrix. The correspondence relationship may be generated so as to satisfy the uniqueness in cooperation with the uniqueness determination unit 27.

  When the control unit 18 issues a start instruction to the first expansion processing execution unit 11 (in the case of translating from Chinese to Japanese via English), the first expansion processing execution unit 11 includes the search unit (extraction unit) 17. The start instruction is given. The search unit (extraction unit) 17 uses a bilingual word dictionary from Chinese to Japanese to extract a Japanese word corresponding to a word constituting the Chinese translation, and from Japanese to English. Processing to extract English words corresponding to the extracted Japanese words using the bilingual word dictionary, and Japanese words corresponding to the extracted English words using the bilingual word dictionary from English to Japanese And processing for extracting words.

  Subsequently, the first expansion processing execution unit 11 finds candidate words that constitute the Chinese translation by combining all the extracted Japanese words while avoiding duplication. The first expansion processing execution unit 11 passes this translation word candidate to the control unit 18. The control unit 18 issues an activation instruction to the hit determination unit 19 and delivers this translated word candidate. The hit determination unit 19 determines whether or not the translation word candidate is included in a word constituting a Japanese sentence. In other words, when the translation candidate is included in the Japanese sentence, the hit determination unit 19 passes the word in the Japanese sentence that matches the translation candidate to the control unit 18 as a determination result.

  The control unit 18 issues an activation instruction to the first corresponding word provisional determination unit 21 and passes this determination result. The first corresponding word provisional determination unit 21 writes the determination result as it is in the correspondence matrix. Therefore, only with this writing process, there may be a plurality of translated words for one word. The uniqueness determination unit 27 determines whether or not there is one translation corresponding to one word based on the result of writing to the correspondence matrix or when writing to the correspondence matrix. The writing to the correspondence matrix is controlled (adjusted) so that one translated word corresponds to one word.

  That is, the uniqueness determination unit 27 may perform the above adjustment each time the writing is performed in cooperation with the first corresponding word tentative determination unit 21, or the first corresponding word tentative determination units 21 to 6. Of the word provisional decision unit 26, the adjustment may be performed by collectively collecting the results of the provisional decision process executed this time (executed expansion process) (corresponding matrixes are prepared as many as the number of expansion logics to be executed) The result of each tentative determination process is written in the matrix, and the uniqueness determination unit 27 generates a new correspondence matrix in which one translated word corresponds to one word based on the writing result to the correspondence matrix).

  In addition, when the control unit 18 issues a start instruction to the second expansion processing execution unit 12 (a case of translating from Japanese to Chinese via English), the second expansion processing execution unit 12 uses a search unit (extraction unit). ) The start instruction is given to 17. The search unit (extraction unit) 17 uses a bilingual word dictionary from Japanese to English to extract an English word corresponding to a word constituting a Japanese sentence, and a bilingual word from English to Japanese. A process of extracting a Japanese word corresponding to the extracted English word using a dictionary is executed.

  Subsequently, the second expansion process execution unit 12 matches the Japanese word extracted in the English word extraction process with the Japanese word extracted in the Japanese word extraction process so as to avoid duplication. Part) 17, using the bilingual word dictionary from Japanese to Chinese, the Chinese word corresponding to the combined Japanese word is extracted as a translation candidate for the word constituting the Japanese sentence. . The second expansion process execution unit 12 passes this candidate word to the control unit 18. The control unit 18 issues an activation instruction to the hit determination unit 19 and delivers this translated word candidate.

  The hit determination unit 19 determines whether the translated word candidate is included in a word constituting a Chinese translated sentence. That is, if the candidate translation is included in the Chinese translation, the hit determination unit 19 passes the word in the Chinese translation that matches the translation candidate to the control unit 18 as a determination result.

  The control unit 18 issues an activation instruction to the second corresponding word provisional determination unit 22 and passes this determination result. The second corresponding word provisional deciding unit 22 writes the determination result as it is in the correspondence matrix or, in cooperation with the uniqueness determining unit 27, writes into the correspondence matrix so that one translated word corresponds to one word. When the second corresponding word tentative determination unit 22 writes the determination result as it is in the correspondence matrix, the uniqueness determination unit 27 is activated after execution of all the temporary determination processes to be executed.

  In addition, when the control unit 18 issues a start instruction to the third expansion processing execution unit 13 (in the case of translating from Chinese to Japanese via a semantic symbol), the third expansion processing execution unit 13 uses the search unit (extraction Part) 17 is instructed to start. The search unit (extraction unit) 17 uses a bilingual word dictionary from Chinese to Japanese to extract a Japanese word corresponding to a word constituting a Chinese translation as a first translation word candidate, Processing to extract semantic symbols of extracted Japanese words using bilingual word dictionary (with semantic symbols) from words to English (other than Japanese or Chinese), and corresponding to the extracted semantic symbols A process of extracting a Japanese word as a second translation word candidate is executed.

  Then, the third expansion processing execution unit 13 combines all the extracted Japanese words while avoiding duplication (by combining the first translation candidate and the second translation candidate while avoiding duplication). A candidate translation of the word is required. The third expansion processing execution unit 13 passes this translation word candidate to the control unit 18. The control unit 18 issues an activation instruction to the hit determination unit 19 and delivers this translated word candidate.

  The hit determination unit 19 determines whether or not the translation word candidate is included in a word constituting a Japanese sentence. In other words, when the translation candidate is included in the Japanese sentence, the hit determination unit 19 passes the word in the Japanese sentence that matches the translation candidate to the control unit 18 as a determination result.

  The control unit 18 issues an activation instruction to the third corresponding word provisional determination unit 23 and passes this determination result. The third corresponding word provisional deciding unit 23 writes the determination result in the corresponding matrix as it is, or, in cooperation with the uniqueness determining unit 27, writes into the corresponding matrix so that one translated word corresponds to one word. When the third corresponding word tentative determination unit 23 writes the determination result as it is in the correspondence matrix, the uniqueness determination unit 27 is activated after execution of all the temporary determination processes to be executed.

  In addition, when the control unit 18 issues an activation instruction to the fourth expansion process execution unit 14 (in the case of translating from Japanese to Chinese via a semantic symbol), the fourth expansion process execution unit 14 uses the search unit (extraction Part) 17 is instructed to start. The search unit (extraction unit) 17 extracts semantic symbols corresponding to words constituting a Japanese sentence using a bilingual word dictionary (with semantic symbols) of Japanese and English (other than Japanese or Chinese). And a process of extracting a Japanese word corresponding to the extracted semantic symbol.

  Then, the fourth expansion process execution unit 14 combines the Japanese words of the semantic symbol extraction process with the Japanese words extracted by the Japanese word extraction process to avoid duplication, and the search unit (extraction unit) 17, using a bilingual word dictionary from Japanese to Chinese, a Chinese word corresponding to the combined Japanese word is extracted as a translation candidate for a word constituting a Japanese sentence. The fourth expansion processing execution unit 14 passes this translation word candidate to the control unit 18. The control unit 18 issues an activation instruction to the hit determination unit 19 and delivers this translated word candidate.

  The hit determination unit 19 determines whether the translated word candidate is included in a word constituting a Chinese translated sentence. That is, if the candidate translation is included in the Chinese translation, the hit determination unit 19 passes the word in the Chinese translation that matches the translation candidate to the control unit 18 as a determination result.

  The control unit 18 issues an activation instruction to the fourth corresponding word provisional determination unit 24 and passes this determination result. The fourth corresponding word provisional deciding unit 24 writes the determination result as it is in the correspondence matrix or, in cooperation with the uniqueness determining unit 27, writes into the correspondence matrix so that one translated word corresponds to one word. When the fourth corresponding word tentative determination unit 24 writes the determination result as it is in the correspondence matrix, the uniqueness determination unit 27 is activated after execution of all the temporary determination processes to be executed.

  When the control unit 18 issues a start instruction to the fifth expansion process execution unit 15 (in the case of translating from Chinese to Japanese via a semantic symbol and English), the fifth expansion process execution unit 15 An activation instruction is issued to the (extraction unit) 17. The search unit (extraction unit) 17 uses a bilingual word dictionary from Chinese to Japanese to extract a Japanese word corresponding to a word constituting a Chinese translation, Using a bilingual English word dictionary (with semantic symbols) (excluding Chinese), processing for extracting the semantic symbols of the extracted Japanese words and extracting English words corresponding to the extracted semantic symbols A process and a process of extracting a semantic symbol corresponding to the extracted English word are executed.

  Then, the fifth expansion processing execution unit 15 combines the semantic symbols extracted corresponding to the Japanese words with the semantic symbols extracted corresponding to the English words, and instructs the search unit (extraction unit) 18 to start up. Put out. The search unit 18 extracts a Japanese word corresponding to the combined meaning symbol as a translation candidate for the word constituting the Chinese translation. The fifth expansion process execution unit 15 passes this candidate word to the control unit 18. The control unit 18 issues an activation instruction to the hit determination unit 19 and delivers this translated word candidate.

  The hit determination unit 19 determines whether or not the translation word candidate is included in a word constituting a Japanese sentence. That is, when the translated word is included in the Japanese sentence, the hit determining unit 19 passes the word in the Japanese sentence that matches the candidate word to the control unit 18 as a determination result.

  The control unit 18 issues an activation instruction to the fifth corresponding word provisional determination unit 25 and passes this determination result. The fifth corresponding word provisional deciding unit 25 writes the determination result as it is in the corresponding matrix or, in cooperation with the uniqueness determining unit 27, writes to the corresponding matrix so that one translated word corresponds to one word. When the fifth corresponding word tentative determination unit 25 writes the determination result as it is in the correspondence matrix, the uniqueness determination unit 27 is activated after execution of all the temporary determination processes to be executed.

  In addition, when the control unit 18 issues a start instruction to the sixth expansion process execution unit 16 (in the case of translating Japanese into Chinese via a semantic symbol and English), the sixth expansion process execution unit 16 An activation instruction is issued to the (extraction unit) 17. The search unit (extraction unit) 17 extracts a semantic symbol of words constituting a Japanese sentence using a bilingual word dictionary (with semantic symbols) of Japanese and English (other than Japanese or Chinese). And a process of extracting an English word corresponding to the extracted semantic symbol and a process of extracting a semantic symbol corresponding to the extracted English word.

  Then, the sixth expansion processing execution unit 16 combines the meaning symbol extracted corresponding to the Japanese word and the meaning symbol extracted corresponding to the English word, and starts the search unit (extraction unit) 17. Give instructions. The search unit (extraction unit) 17 uses a process of extracting a Japanese word corresponding to the combined semantic symbol and a Japanese-to-Chinese bilingual word dictionary to correspond to the extracted Japanese word. A process of extracting a Chinese word as a translation candidate of a word constituting a Japanese sentence is executed. The sixth expansion process execution unit 16 passes the translated word candidate to the control unit 18. The control unit 18 issues an activation instruction to the hit determination unit 19 and delivers this translated word candidate.

  The hit determination unit 19 determines whether the translated word candidate is included in a word constituting a Chinese translated sentence. That is, if the candidate translation is included in the Chinese translation, the hit determination unit 19 passes the word in the Chinese translation that matches the translation candidate to the control unit 18 as a determination result.

  The control unit 18 issues an activation instruction to the sixth corresponding word provisional determination unit 26 and passes this determination result. The sixth corresponding word provisional deciding unit 26 writes the determination result as it is in the correspondence matrix or, in cooperation with the uniqueness determining unit 27, writes into the correspondence matrix so that one translated word corresponds to one word. When the sixth corresponding word tentative determination unit 26 writes the determination result as it is in the correspondence matrix, the uniqueness determination unit 27 is activated after execution of all the temporary determination processes to be executed.

In FIG. 1A, the expansion process of the correspondence relationship between two or more words is specified and executed, but at least one or more expansion processes may be specified and executed.
FIG. 2 is a diagram for explaining the operation of the inter-word correspondence determination processing unit of the present embodiment.

  As shown in FIG. 2, first, a sentence / translation pair, for example, a Japanese sentence 1010 (“good partnership” in the figure) and a Chinese character are translated by a user from a Japanese-Chinese bilingual example dictionary. A translated sentence 1020 is designated. Then, by executing an existing morphological analysis on the specified sentence / translation sentence pair, the sentence 1010 and the translation sentence 1020 are divided into words constituting the sentence / translation sentence. These divided words are stored in, for example, a one-dimensional list LS1 for those constituting a Japanese sentence, and for those constituting a Chinese translation, for example, It is stored in the list LS2 which is a one-dimensional array.

  In FIG. 2, as a result of the morphological analysis, the sentence 1010, that is, “good alliance” is divided into a word 1011 (“good”), a word 1012 (“association”), and a word 1013 (“relation”). The The translated sentence 1020 is divided into a Chinese word 1021, a word 1022, and a word 1023.

  The present invention relates to a technique for associating a sentence previously associated as a bilingual example dictionary with a plurality of words constituting the translated sentence (each example), for example, a word that is not associated with the bilingual word dictionary. About. In FIG. 2, correspondence between some words, that is, the Japanese word 1013 (“relation”) and the Chinese word 1023 are assigned by a bilingual word dictionary between the first language and the second language. It has been.

An example of the corresponding word (candidate) expansion process performed prior to the process for determining the correspondence between words according to the present embodiment will be further described with reference to FIG.
That is, in the figure, correspondence between Japanese words (word 1011 (“good”), word 1012 (“affiliation”)) and correspondence between Chinese words (word 1021, word 1022) that are not in correspondence. For example, using a bilingual word dictionary that associates the correspondence between Japanese words to English words, the English words corresponding to the Japanese words are extracted, and the Japanese words are derived from the English words. Extracting Japanese words corresponding to the extracted English words using a bilingual word dictionary that assigns correspondences to the words, that is, reversely extracting Japanese words from the extracted English words Thus, for example, it is possible to expand the range of Japanese words that should be candidates for correspondence with Chinese words. As a result, the range of original Japanese words (candidates) corresponding to Chinese words can be handled even if a bilingual word dictionary that provides correspondence between Japanese and Chinese words is used. Even if a Chinese word does not exist, there may be a corresponding Chinese word for a certain word included in the expanded Japanese word (candidate) range corresponding to the Chinese word. Arise. Therefore, it is possible to expand and add correspondences between a plurality of words constituting a Japanese sentence and a Chinese translation (attach to a wider range of words).

FIG. 3 is a flowchart illustrating an example of a process for determining the correspondence between words according to the present embodiment. This flowchart is executed by, for example, the correspondence determination processing unit 10 in FIG. 1A.
In FIG. 3, first, in step S101, a plurality of words constituting a Japanese sentence are obtained as a result of the morphological analysis described above, and the plurality of words are stored in a list LS1 which is a one-dimensional array. A plurality of words constituting a Chinese translation are obtained, and the plurality of words are stored in a list LS2 which is a one-dimensional array. In the following step S102, Chinese words are extracted one by one from the list LS2, and in step S103, a process (part 1) for expanding the range of Japanese words (candidates) corresponding to the extracted words is shown in FIG. This is executed by the first expansion processing execution unit 11. In step S104, a hit determination is made as to whether the expanded word (candidate) range is included in the words constituting the example, and writing to the above-described correspondence matrix is performed.

  In step S105, Japanese words are extracted one by one from the list LS1, and in step S106, a process (part 1) for expanding the range of Chinese words (candidates) corresponding to the extracted words is shown. It is executed by the 1A second expansion processing execution unit 12. In step S107, a hit determination is made as to whether the expanded word (candidate) range is included in the words constituting the example, and writing to the above-described correspondence matrix is performed.

  In step S108, Chinese words are extracted from the list LS2 one by one. In step S109, a process (part 2) for expanding the range of Japanese words (candidates) corresponding to the extracted words is shown. It is executed by the 1A third expansion processing execution unit 13. In step S110, a hit determination is made as to whether the expanded word (candidate) range is included in the words constituting the example, and writing to the above-described correspondence matrix is performed.

  In step S111, Japanese words are extracted one by one from the list LS1, and in step S112, a process (part 2) for expanding the range of Chinese words (candidates) corresponding to the extracted words is shown. It is executed by the first expansion processing execution unit 14 of 1A. Then, in step S113, hit determination is performed as to whether the expanded word (candidate) range is included in the words constituting the example, and writing to the above-described correspondence matrix is performed.

  In step S114, Chinese words are extracted from the list LS2 one by one. In step S115, a process (part 3) for expanding the range of Japanese words (candidates) corresponding to the extracted words is shown. It is executed by the first expansion processing execution unit 15 of 1A. In step S116, a hit determination is made as to whether the expanded word (candidate) range is included in the words constituting the example, and writing to the above-described correspondence matrix is performed.

  In step S117, Japanese words are extracted one by one from the list LS1, and in step S118, a process (part 3) for expanding the range of Chinese words (candidates) corresponding to the extracted words is shown. It is executed by the first expansion processing execution unit 16 of 1A. In step S119, a hit determination is made as to whether the expanded word (candidate) range is included in the words constituting the example, and writing to the above-described correspondence matrix is performed.

  Note that the user interface uses the information of the correspondence matrix to correspond to the other language on the screen when the position of a word in one language displayed on the screen is pointed by means such as a mouse. Change the display color of the word.

  In addition, in the relationship with the block diagram of FIG. 1, the case where all the processes by the 1st expansion process execution part-the 6th expansion process execution part are performed corresponds to the flowchart of FIG. When there are a large number of words constituting each sentence / translation of a specified sentence / translation pair, usually, by applying the bilingual word dictionary, there is no correspondence between many words.

  In such a case, as shown in the flowchart of FIG. 3, if as many expansion processes (enlargement logic) as possible are executed and the results of the processes are combined, correspondence between words can be achieved with any expansion logic. As a result, it is possible to increase the number of word pairs that can be associated with each other.

FIG. 4 is a correspondence matrix corresponding to the Japanese and Chinese example sentences in FIG.
In the example of FIG. 4, the first line of the correspondence matrix is the Japanese word 1011 (“good”), the second line is the Japanese word 1012 (“affiliation”), and the third line is the Japanese word. 1013 ("Relation"). The first column of the correspondence matrix corresponds to the Chinese word 1021, the second column corresponds to the Chinese word 1022, and the third column corresponds to the Chinese word 1023. That is, in this case, the presence of a mark in the i-th row and j-th column of the correspondence matrix indicates that the correspondence between the i-th word in Japanese and the j-th word in Chinese is originally or as a result of the expansion process. It shows that it was attached.

  FIG. 5 is a flowchart showing in more detail the writing process to the corresponding matrix executed in each step shown as steps S104, S107, S110, S113, S116, and S119 in FIG. In starting the processing of the flowchart of FIG. 5, as a result of the expansion processing by the previous expansion processing unit or the determination processing by the hit determination unit 19, the i-th word in Japanese and the j-th word in Chinese It is assumed that a correspondence relationship is attached.

  In FIG. 5, first, in step S201, it is determined whether a mark exists in any column of the i row of the correspondence matrix. In this determination, when it is determined that a mark is present in any column of the i row, a series of processing ends. On the other hand, if there is no mark in any column of the i row, the process proceeds to step S202. In step S202, it is determined whether a mark exists in any row of the j column of the correspondence matrix. In this determination, if it is determined that a mark is present in any row of the j column, a series of processing ends. On the other hand, if there is no mark in any row of the j column, the process proceeds to step S203, and the mark is written in the i row and j column of the correspondence matrix.

FIG. 6 is a diagram for explaining processing (part 1) for expanding the range of Japanese words (candidates) corresponding to Chinese words.
In FIG. 6, the fact that the Japanese word 1031 (“excellent”) corresponds to the Chinese word 1030 is indicated by a bold arrow starting from the Chinese word 1030. Also, the Japanese word 1031 (“excellent”), the English word (word 1032 ₁ (“excel”), word 1032 ₂ (“excellent”), word 1032 ₃ (“great”)), word 1032 ₄ ("Strong") and the word 1032 ₅ ("superior")) correspond to each other by an arrow starting from the word 1031. Furthermore, they word 1032 _1, 1032 _2, 1032 _3, 1032 _4, 1032 ₅ by the arrows starting from the their word 1032 _1, 1032 _2, 1032 _3, 1032 _4, 1032 ₅ and the Japanese word (word 1033 ₁ (“Excellent”), word 1033 ₂ (“good”), word 1033 ₃ (“win”), word 1033 ₄ (“excellent”), word 1033 ₅ (“excellent”), word 1033 ₆ (“rugged”) , Word 1033 ₇ (“durable”), word 1033 ₈ (“advanced”), word 1033 ₉ (“dominant”)) are shown to correspond.

That is, at first, even if a bilingual word dictionary between Japanese and Chinese was used, only the Japanese word 1031 (“excellent”) was supported for the Chinese word 1030. As a result of the processing, the Japanese words 1033 _{1 to} 1033 ₉ can also correspond to the Chinese word 1030. That is, the corresponding word (candidate) has been expanded.

  FIG. 7 is a flowchart of the process (part 1) for expanding the range of Japanese words (candidates) corresponding to Chinese words. This flowchart is executed by the first expansion processing execution unit 11 of FIG.

  In FIG. 7, first, in step S301, the search unit 17 uses a bilingual word dictionary from Chinese to Japanese (Chunichi bilingual dictionary) to correspond to one Chinese word extracted from the list LS2. Japanese words are searched. The list LS2 is a one-dimensional array in which a plurality of words constituting a specified Chinese example sentence (example) are stored.

  Subsequently, in step S302, the Japanese words of the search result of step S301, that is, the words written in Japanese are stored in lists (one-dimensional arrays) L11 and L12, respectively. In step S303, it is determined whether a word (Japanese notation) exists in the list L11.

  If it is determined in step S303 that the word (in Japanese notation) does not exist in the list L11, the series of processing ends. On the other hand, if it is determined in step S303 that a word (Japanese notation) is present in the list L11, the process proceeds to step S304, the first word (Japanese notation) in the list L11 is extracted, stored in the buffer Temp1, The word (data) stored in Temp1 is deleted from the list L11. In step S305, using a bilingual word dictionary from Japanese to English (Japanese-English bilingual dictionary), a word (English notation) corresponding to the word (Japanese notation) stored in Temp1 is searched, and step S306 is performed. In step S305, the search result word (in English) is stored in the list L3.

  In subsequent step S307, it is determined whether or not an English word (data) is present in the list L3. If it is determined in step S307 that the word (English notation) does not exist in the list L3, the process returns to step S303. On the other hand, if it is determined in step S307 that the word (English notation) is present in the list L3, the process proceeds to step S308, the first word (English notation) in the list L3 is extracted, stored in the buffer Temp2, and the Temp2 Is deleted from the list L3. In step S309, a word (Japanese notation) corresponding to the word (English notation) stored in Temp2 is searched using an English-to-Japanese bilingual word dictionary (English-Japanese bilingual dictionary), and step S310. Thus, the word (Japanese notation) of the search result in step S309 is added (stored) to the list L12. Then, the process returns to step S307.

Through such a series of processing, the list L12 includes a range of Japanese words (candidates) corresponding to one Chinese word (for example, the word 1030 in FIG. 6) (for example, the word 1031 in FIG. 6, 1033 _{1 to} 1033 ₉ ) are acquired.

FIG. 8 is a diagram illustrating a process (part 1) for expanding the range of Chinese words (candidates) corresponding to Japanese words.
In FIG. 8, a Japanese word 1040 (“excellent”) is replaced with an English word (word 1042 ₁ (“excel”), word 1042 ₂ (“excellence”), word 1042 ₃ (“superior”), word 1042 _4. ("Transcend"), corresponding to the word 1042 ₅ ("sublime") is indicated by an arrow starting from the Japanese word 1040. In addition, by the arrow whose starting point is those English words 1042 _{1-1042 9,} those English words 1042 _{1-1042 9,} the Japanese word (word 1043 ₁ ( "excellent"), the words 1043 ₂ ( "excellence ”), Word 1043 ₃ (“ win ”), word 1043 ₄ (“ excellent ”), word 1043 ₅ (“ good ”), word 1043 ₆ (“ dominant ”), word 1043 ₇ (“ advanced ”), word 1043 ₈ (“transcendence”) and the word 1043 ₉ (“Takao”)). In addition, the word of the original Japanese) 1040 ( "excellence"), the words 1040,1043 _{1-1043 9} a combination of the words 1043 _{1-1043 9} of the added Japanese over the English, the words of Chinese Corresponding to 1041 ₁ , 1041 ₂ , 1044 ₁ -1044 ₆ are indicated by arrows starting from these words 1040, 1043 ₁ -1043 ₉ . The fact that the Chinese words corresponding to the Japanese words 1043 ₂ , 1043 ₄ , 1043 ₈ , and 1043 ₉ do not exist in the bilingual word dictionary from Japanese to Chinese is indicated by “x” in FIG. It is shown.

That is, at first, even if a word dictionary between Japanese and Chinese was used, only the Chinese words 1041 ₁ and 1041 ₂ were supported for the Japanese word 1040 (“excellence”). by the expansion process, the further to the Japanese word 1040 ( "excellence"), also the words 1044 _{1-1044 6} of the Chinese began to corresponding stick. That is, the corresponding word (candidate) has been expanded.

  FIG. 9 is a flowchart of the process (part 1) for expanding the range of Chinese words (candidates) corresponding to Japanese words. This flowchart is executed by the second expansion processing execution unit 12 of FIG. 1A.

  In FIG. 9, first, in step S401, using a bilingual word dictionary from Japanese to English (Japanese-English bilingual dictionary), one Japanese word extracted from the list LS1 is stored in the list (one-dimensional array) L1. While storing, the search unit 17 searches for an English word corresponding to the extracted word. The list LS1 is a one-dimensional array in which a plurality of words constituting a designated Japanese example sentence (example) are stored.

  Subsequently, in step S402, the English word of the search result of step S401, that is, the word written in English is stored in the list (one-dimensional array) L3. In step S403, it is determined whether a word (in English) is present in the list L3.

  If it is determined in step S403 that the word (English notation) does not exist in the list L3, the control moves to the subsequent step S407. On the other hand, if it is determined in step S403 that the word (English notation) is present in the list L3, the process proceeds to step S404, the first word (English notation) in the list L3 is extracted, stored in the buffer Temp1, and the Temp1 Is deleted from the list L3. In step S405, a word (Japanese notation) corresponding to the word (English notation) stored in Temp1 is searched using an English-to-Japanese parallel word dictionary (English-Japanese parallel translation dictionary), and step S406. Then, the word (Japanese notation) of the search result in step S405 is added (stored) to the list L1.

  In step S407, it is determined whether or not a word (data) written in Japanese exists in the list L1. If it is determined in step S407 that the word (in Japanese notation) does not exist in the list L1, the series of processing ends. On the other hand, if it is determined in step S407 that a word (in Japanese notation) is present in the list L1, the process proceeds to step S408, and the first word (in Japanese notation) in the list L1 is extracted and stored in the buffer Temp2. The word (data) stored in Temp2 is deleted from the list L1. In step S409, a word (Chinese notation) corresponding to the word (Japanese notation) stored in Temp2 is searched using a Japanese-to-Chinese bilingual word dictionary (Japanese-Chinese bilingual dictionary), In step S410, the search result word (in Chinese notation) in step S409 is added (stored) to the list L2. Then, the process returns to step S407.

Through such a series of processes, the list L2 includes a range of Chinese words (candidates) corresponding to one Japanese word (for example, the word 1040 in FIG. 8) (for example, the word 1041 _{1 in} FIG. 8). , 1041 ₂ , 1044 _{1 to} 1044 ₆ ).

FIG. 10 is a diagram for explaining the process (part 2) for expanding the range of Japanese words (candidates) corresponding to Chinese words.
In FIG. 10, a Japanese word 1051 (“decline”) corresponds to a Chinese word 1050 by a bold arrow starting from the Chinese word 1050. Also, the Japanese word 1051 (“drop”), a semantic symbol (semantic symbol 1052 ₁ (“* fall *”), a semantic symbol 1052 ₂ (“* drop *”), a semantic symbol 1052 ₃ (“* decrease”) Corresponding to “*”)) is indicated by an arrow starting from the word 1051. Furthermore, the arrow that starts them meaning symbols 1052 _1, 1052 _2, 1052 _3, they mean symbol 1052 _1, 1052 _2, 1052 ₃ and Japanese words (word 1053 ₁ ( "down"), the word 1053 ₂ ( It is shown that the words “falling”), word 1053 ₃ (“falling”), and word 1053 ₄ (“falling”) correspond to each other.

That is, at first, even if a bilingual word dictionary between Japanese and Chinese was used, only the Japanese word 1051 (“decline”) was associated with the Chinese word 1050. As a result, the Japanese words 1053 _{1 to} 1053 ₄ are also associated with the Chinese word 1050. That is, the corresponding word (candidate) has been expanded.

  FIG. 11 is a flowchart of the process (part 2) for expanding the range of Japanese words (candidates) corresponding to Chinese words. This flowchart is executed by the third expansion processing execution unit 13 in FIG. 1A.

  In FIG. 11, first, in step S501, a Japanese corresponding to one Chinese word extracted from the list LS2 by the search unit 17 using a bilingual dictionary of Chinese and Japanese (Chinese bilingual dictionary). The word of the word is searched.

  In subsequent step S502, the Japanese words of the search result of step S501, that is, the words in Japanese notation are stored in lists (one-dimensional arrays) L11 and L12, respectively. In step S503, it is determined whether a word (Japanese notation) exists in the list L11.

  If it is determined in step S503 that the word (Japanese notation) does not exist in the list L11, the series of processing ends. On the other hand, if it is determined in step S503 that a word (in Japanese notation) is present in the list L11, the process proceeds to step S504, where the first word (in Japanese notation) in the list L11 is extracted and stored in the buffer Temp1. The word (data) stored in Temp1 is deleted from the list L11. Then, in step S505, a Japanese-English bilingual word dictionary (with semantic symbols) is used to search for a semantic symbol corresponding to the word (Japanese notation) stored in Temp1, and in step S506, in step S505. The meaning symbol of the search result is stored in the list L1M.

  In subsequent step S507, it is determined whether or not there is a semantic symbol (data) in the list L1M. If it is determined in step S507 that no semantic symbol exists in the list L1M, the process returns to step S503. On the other hand, if it is determined in step S507 that a semantic symbol is present in the list L1M, the process proceeds to step S508, where the first semantic symbol of the list L1M is extracted, stored in the buffer Temp2, and the semantic symbol stored in Temp2 ( Data) is deleted from the list L1M. In step S509, a word (in Japanese notation) corresponding to the semantic symbol stored in Temp2 is searched using a bilingual dictionary of Japanese and English (with semantic symbols), and in step S510, in step S509. The search result word (in Japanese) is added (stored) to the list L12. Then, the process returns to step S507.

Through such a series of processes, the list L12 includes a range of Japanese words (candidates) corresponding to one Chinese word (for example, the word 1050 in FIG. 10) (for example, the word 1051, FIG. 10). 1053 _{1 to} 1053 ₄ ) are acquired.

FIG. 12 is a diagram for explaining a process (part 2) for expanding the range of Chinese words (candidates) corresponding to Japanese words.
In FIG. 12, a Japanese word 1060 (“down”) corresponds to a semantic symbol (semantic symbol 1062 ₁ (“* descend *”), semantic symbol 1062 ₂ (“* fall *”)). It is indicated by an arrow starting from the word 1060. Further, the arrow that starts them meaning symbols 1062 _{1-1062 2,} they mean symbol 1062 _{1-1062 2,} the Japanese word (word 1063 ₁ ( "drop"), a word 1063 ₂ ( "fall"), The word 1063 ₃ (“decline”) and the word 1063 ₄ (“decline”)) are shown to correspond. In addition, the original and the Japanese word 1060 ( "down"), meaning the words of the symbols have been added via the Japanese 1063 _{1-1063 4} and the word 1060,1063 _{1-1063 4} combined is, of the second language correspond to words 1061,1064 _{1-1064 4} is indicated by an arrow that starts them words 1060,1063 _{1-1063 4.}

That is, at first, even if a Japanese-Chinese bilingual word dictionary was used, only the Chinese word 1061 was supported for the Japanese word 1060 (“down”). As a result, the Chinese words 1064 _{1 to} 10644 ₄ can also correspond to the Japanese word 1060 (“down”). That is, the corresponding word (candidate) has been expanded.

  FIG. 13 is a flowchart of the process (part 2) for expanding the range of Chinese words (candidates) corresponding to Japanese words. This flowchart is executed by the fourth expansion processing execution unit 14 of FIG. 1A.

  In FIG. 13, first, in step S601, one Japanese word extracted from the list LS1 is stored in the list (one-dimensional array) L1 using a Japanese-English bilingual dictionary (with semantic symbols). At the same time, the retrieval unit 17 retrieves a semantic symbol corresponding to the extracted word.

  In subsequent step S602, the semantic symbols of the search result in step S601 are stored in the list (one-dimensional array) L1M. In step S603, it is determined whether there is a semantic symbol in the list L1M.

  If it is determined in step S603 that there is no semantic symbol in the list L1M, the control moves to the subsequent step S607. On the other hand, if it is determined in step S603 that the semantic symbol is present in the list L1M, the process proceeds to step S604, where the leading semantic symbol of the list L1M is extracted, stored in the buffer Temp1, and stored in the Temp1 ( Data) is deleted from the list L1M. In step S605, a word (in Japanese) corresponding to the semantic symbol stored in Temp1 is searched using a Japanese and English word dictionary (with semantic symbols). In step S606, the search in step S605 is performed. The resulting word (in Japanese) is added (stored) to the list L1.

  In step S607, it is determined whether or not a word (data) written in Japanese exists in the list L1. If it is determined in step S607 that the word (in Japanese notation) does not exist in the list L1, the series of processing ends. On the other hand, if it is determined in step S607 that a word (in Japanese notation) is present in the list L1, the process proceeds to step S608, where the first word (in Japanese notation) in the list L1 is extracted and stored in the buffer Temp2. The word (data) stored in Temp2 is deleted from the list L1. In step S609, a word (Chinese notation) corresponding to the word (Japanese notation) stored in Temp2 is searched using a Japanese-to-Chinese bilingual word dictionary (Japanese-Chinese bilingual dictionary), In step S610, the search result word (in Chinese notation) in step S609 is added (stored) to the list L2. Then, the process returns to step S607.

Through such a series of processing, the list L2 includes a range of Chinese words (candidates) corresponding to one Japanese word (for example, word 1060 in FIG. 12) (for example, word 1061, 1064 _{1 to} 1064 ₄ ) are acquired.

FIG. 14 is a diagram for explaining the process (part 3) for expanding the range of Japanese words (candidates) corresponding to Chinese words.
In FIG. 14, the fact that the Japanese word 1071 (“decline”) corresponds to the Chinese word 1070 is indicated by a bold arrow starting from the Chinese word 1070. Also, the Japanese word 1071 (“drop”), a semantic symbol (semantic symbol 1072 ₁ (“* fall *”), a semantic symbol 1072 ₂ (“* drop *”), a semantic symbol 1072 ₃ (“* decrease”). Corresponding to “*”)) is indicated by an arrow starting from the word 1071. Furthermore, the arrow that starts them meaning symbols 1072 _1, 1072 _2, 1072 _3, they mean symbol 1072 _1, 1072 _2, 1072 ₃ and English words (word 1073 ₁ ( "fall"), the word 1073 ₂ ( " drop ") and the word 1073 ₃ (" decrease ")) are shown to correspond.

In addition, by the arrow whose starting point is those English words 1073 _1, 1073 _2, 1073 _3, the words of their English 1073 _1, 1073 _2, 1073 ₃ the meaning symbol (meaning symbol 1074 ₁ ( "* fall2 *"), meaning Symbol 1074 ₂ (“* fall3 *”), semantic symbol 1074 ₃ (“* fall4 *”), semantic symbol 1074 ₄ (“* drop2 *”), semantic symbol 1074 ₅ (“* drop3 *”)) correspond to It has been shown.

The meaning symbols 1072 _{1 to} 1072 ₃ and the meaning symbols 1074 _{1 to} 1074 ₅ are combined, that is, the meaning symbols 1072 _{1 to} 1072 ₃ , 1074 _{1 to} 1074 ₅ are indicated by the arrows. 1072 _{1 to} 1072 ₃ , 1074 _{1 to} 1074 ₅ , and Japanese words (word 1075 ₁ (“down”), word 1075 ₂ (“down”), word 1075 ₃ (“down”), word 1075 ₄ (“ Descent "), word 1075 ₅ (" falling "), word 1075 ₆ (" falling castle "), word 1075 ₇ (" autumn "), word 1075 ₈ (" drop "), word 1075 ₉ (" dropping ")) Corresponding is shown.

That is, at first, even if a bilingual word dictionary of Japanese and Chinese was used, only the Japanese word 1071 (“decline”) was associated with the Chinese word 1070. As a result, the Japanese words 1075 _{1 to} 1075 ₉ can also correspond to the Chinese word 1070. That is, the corresponding word (candidate) has been expanded.

  FIG. 15 is a flowchart of the process (part 3) for expanding the range of Japanese words (candidates) corresponding to Chinese words. This flowchart is executed by the fifth expansion processing execution unit 15 in FIG. 1A.

  In FIG. 15, first, in step S701, using the bilingual word dictionary from Chinese to Japanese (Chinese-Japanese bilingual dictionary), the search unit 17 corresponds to one Chinese word extracted from the list LS2. Japanese words are searched.

  In subsequent step S702, the Japanese words of the search results in step S701, that is, the words in Japanese notation are stored in lists (one-dimensional arrays) L11 and L12, respectively. In step S703, it is determined whether a word (Japanese notation) exists in the list L11.

If it is determined in step S703 that the word (Japanese notation) does not exist in the list L11, the control moves to step S703. On the other hand, if it is determined in step S703 that a word (in Japanese notation) is present in the list L11, the process proceeds to step S704, and the first word (in Japanese notation) in the list L11 is extracted and stored in the buffer Temp1. The word (data) stored in Temp1 is deleted from the list L11. In step S705, a Japanese-English bilingual word dictionary (with semantic symbols) is used to search for a semantic symbol corresponding to the word (Japanese notation) stored in Temp1, and in step S706, the processing in step S705 is performed. The meaning symbols of the search results are stored in the lists L1M ₁ and L1M ₂ .

In the subsequent step S707, the duplicates of the lists L1M ₁ and L1M ₂ are removed. As a result, the same semantic symbols are not duplicated in the lists L1M ₁ and L1M ₂ . Then, in step S 708, or refers to a list L1M ₁ symbol (data) exists is determined.

In step S 708, if the meaning symbol is determined to exist in the list L1M _1, the process proceeds to step S709, extracts the first meaning symbols of the list L1M _1, meaning symbols stored in the buffer Temp2, stored in its Temp2 ( to delete the data) from the top list L1M _1. In step S710, a word (English notation) corresponding to the semantic symbol stored in Temp2 is searched using a Japanese-English bilingual dictionary (with semantic symbols). In step S711, the search in step S710 is performed. The resulting word (in English) is added (stored) to the list L3. Then, the process returns to step S708.

On the other hand, in step S 708, if the meaning symbol is determined not to exist in the list L1M _1, the process proceeds to step S712, duplicates the list L3 is removed. As a result, the same English word is not duplicated in the list L3. In subsequent step S713, it is determined whether or not a word (English notation) exists in the list L3.

  If it is determined in step S713 that a word (English notation) is present in the list L3, the process proceeds to step S714, the first word (English notation) in the list L3 is extracted, stored in the buffer Temp3, and stored in that Temp3. The deleted word (data) is deleted from the list L3. In step S715, a Japanese-English bilingual word dictionary (with semantic symbols) is used to search for a semantic symbol corresponding to the word (English notation) stored in Temp3. In step S716, the search in step S715 is performed. The resulting semantic symbols are stored in the list L3M.

On the other hand, if it is determined in step S713 that the word (English notation) does not exist in the list L3, the process proceeds to step S717, and the duplicate of the list L3M is removed (as a result, the same semantic symbol is duplicated in the list L3M). with it is no longer) to be present and, to add to avoid duplicate the contents of the list L1M ₂ to list L3M. In subsequent step S718, it is determined whether there is a semantic symbol (data) in the list L3M.

  If it is determined in step S718 that a semantic symbol is present in the list L3M, the process proceeds to step S719, the first semantic symbol of the list L3M is extracted, stored in the buffer Temp4, and the semantic symbol (data) stored in the Temp4. Is deleted from the list L3M. In step S720, a Japanese-English bilingual dictionary (with semantic symbols) is used to search for a word (in Japanese notation) corresponding to the semantic symbol stored in Temp4. In step S721, the processing in step S720 is performed. The search result word (in Japanese) is added (stored) to the list L12. Then, the process returns to step S718.

On the other hand, if it is determined in step S718 that the semantic symbol does not exist in the list L3M, the series of processing ends.
Through such a series of processing, the list L12 includes a range of Japanese words (candidates) corresponding to one Chinese word (for example, the word 1070 in FIG. 14) (for example, the word 1071, FIG. 14). 1075 _{1 to} 1075 ₉ ) are acquired.

FIG. 16 is a diagram for explaining a process (part 3) for expanding the range of Chinese words (candidates) corresponding to Japanese words.
In FIG. 16, it is the word 1080 that the Japanese word 1080 (“down”) corresponds to the semantic symbol (the semantic symbol 1081 ₁ (“* descend *”)) and the semantic symbol 1081 ₂ (“* fall *”). Furthermore, the meaning symbols 1081 ₁ , 1081 ₂ and the meaning symbols 1081 ₁ , 1081 ₂ and English words (words 1082 ₁ (“descend”), The word 1082 ₂ (“fall”) is shown to correspond.

The English words 1082 ₁ , 1082 ₂ start from the arrows, and the English words 1082 ₁ , 1082 ₂ , meaning symbols (semantic symbols 1083 ₁ (“* descend2 *”)), semantic symbols 1083 ₂ (“* fall2 * ”), semantic symbols 1083 ₃ (“ * fall3 * ”), and semantic symbols 1083 ₄ (“ * fall4 * ”).

The meaning symbols 1081 _{1 to} 1081 ₂ and the meaning symbols 1083 _{1 to} 1083 ₄ are combined, that is, the meaning symbols are represented by arrows starting from the meaning symbols 1081 _{1 to} 1081 ₂ and 1083 _{1 to} 1083 _4. 1081 _{1 to} 1081 ₂ , 1083 _{1 to} 1083 ₄ , Japanese words (word 1084 ₁ (“fall”), word 1084 ₂ (“fall”), word 1084 ₃ (“fall”)), word 1084 ₄ (“ The word 1084 ₅ (“fall”), the word 1084 ₆ (“falling”), the word 1084 ₇ (“falling castle”), and the word 1084 ₈ (“autumn”) are shown to correspond.

Then, a Japanese word 1080, word obtained by combining the word 1084 _{1-1084 8,} i.e., by the arrows starting from the Japanese word 1080,1084 _{1-1084 8,} those words Japanese 1080,1084 _{1 to} 1084 ₈ and the Chinese words 1085 and 1086 _{1 to} 1086 ₈ are shown to correspond.

That is, at first, even if a bilingual word dictionary of Japanese and Chinese was used, only the Chinese word 1085 was supported for the Japanese word 1080 (“down”). As a result, the Chinese word 1086 _{1 to} 10886 ₈ can also correspond to the Japanese word 1080 (“down”). That is, the corresponding word (candidate) has been expanded.

  FIG. 17 is a flowchart of the process (part 3) for expanding the range of Chinese words (candidates) corresponding to Japanese words. This flowchart is executed by the sixth expansion processing execution unit 16 in FIG. 1A.

  In FIG. 17, first, in step S801, a semantic symbol corresponding to one Japanese word extracted from the list LS1 is retrieved by the search unit 17 using a Japanese-English bilingual dictionary (with semantic symbols). Searched.

In step S802, the meaning symbols of the search results in step S801 are stored in lists (one-dimensional arrays) L1M ₁ and L1M ₂ , respectively. Then, in step S803, whether there is meant the symbols in the list L1M ₁ is determined.

In step S803, if the meaning symbol is determined to exist in the list L1M _1, the process proceeds to step S804, the extracts the first meaning symbols of the list L1M _1, stored in the buffer Temp1, the meaning symbols stored in the Temp1 be deleted from the list L1M _1. In step S805, an English word corresponding to the semantic symbol stored in Temp1 is searched using a Japanese-English bilingual dictionary (with semantic symbols). In step S806, the search result of step S805 is searched. A word (in English) is added (stored) to the list L3.

On the other hand, in step S803, if the meaning symbol is determined not to exist in the list L1M _1, the flow advances to step S807. In step S807, the duplication of the list L3 is removed. As a result, the same English word is not duplicated in the list L3. In step S808, it is determined whether there is an English word (data) in the list L3.

  If it is determined in step S808 that the word (English notation) is present in the list L3, the process proceeds to step S809, the first word (English notation) in the list L3 is extracted, stored in the buffer Temp2, and stored in the Temp2 The deleted word (data) is deleted from the list L3. In step S810, a single meaning symbol corresponding to the word (in English notation) stored in Temp2 is searched using a Japanese-English bilingual dictionary (with semantic symbols), and in step S811, step S810 is performed. The meaning symbol of the search result is added (stored) to the list L3M. Then, the process returns to step S808.

  On the other hand, if it is determined in step S808 that the word (English notation) does not exist in the list L3, the process proceeds to step S812, and the duplicate of the list L3M is removed (as a result, the same semantic symbol is duplicated in the list L3M). And the contents of the list L1M2 are added to the list L3M while avoiding duplication. In subsequent step S813, it is determined whether there is a semantic symbol in the list L3M.

  If it is determined in step S813 that a semantic symbol exists in the list L3M, the process proceeds to step S814, the first semantic symbol of the list L3M is extracted, stored in the buffer Temp3, and the semantic symbol stored in the Temp3 is stored in the list L3M. Delete from above. In step S815, a Japanese word corresponding to the semantic symbol stored in Temp3 is searched using a Japanese-English bilingual dictionary (with semantic symbols). In step S816, the search result of step S815 is searched. Are added (stored) to the list L1.

  On the other hand, if it is determined in step S813 that the semantic symbol does not exist in the list L3M, the process proceeds to step S817, and the duplication of the list L1 is removed. As a result, the same Japanese word is not duplicated in the list L1. In subsequent step S818, it is determined whether or not a word (data) written in Japanese exists in the list L1.

  If it is determined in step S818 that a word (Japanese notation) is present in the list L1, the process proceeds to step S819, the first word (Japanese notation) in the list L1 is extracted, stored in the buffer Temp4, and the Temp4 Is deleted from the list L1. In step S820, a word (Chinese notation) corresponding to the word (Japanese notation) stored in Temp4 is searched using a Japanese-to-Chinese bilingual word dictionary (Japanese-Chinese bilingual dictionary). In step S821, the search result word (in Chinese notation) in step S820 is added (stored) to the list L2. Then, the process returns to step S818.

On the other hand, if it is determined in step S818 that the word (Japanese notation) does not exist in the list L1, the series of processing ends.
Through such a series of processing, the list L2 includes a range of Chinese words (candidates) corresponding to one Japanese word (for example, the word 1080 in FIG. 16) (for example, the word 1085 in FIG. 16, 1086 _{1 to} 1086 ₈ ) are acquired.

Note that the processing for expanding the correspondence between words according to the present embodiment described above can also be applied in the field of machine translation.
FIG. 18 is a diagram for explaining a case where the processing for expanding the correspondence between words according to the present embodiment is applied to machine translation (translation processing).

In FIG. 18, even if a Japanese-to-Chinese bilingual word dictionary (Japanese-Chinese bilingual dictionary) is used for a Japanese word 1090 (“Desktop”), a corresponding Chinese word may not be found. Assumed. In such a case, for example, the English word 1091 (“excel”) corresponding to the Japanese word 1090 (“exclusion”) is extracted by the expansion processing by the second expansion processing unit of this embodiment, and the English Japanese words (word 1092 ₁ (“excellent”), word 1092 ₂ (“excellent”), word 1092 ₃ (“excellent”)) corresponding to the word 1091 are further extracted. Then, for the extracted Japanese words 1092 ₁ , 1092 ₂ , and 1092 ₃ , a corresponding Chinese word 1093 _{1 is obtained} by using a Japanese-to-Chinese bilingual word dictionary (Japanese-Chinese bilingual dictionary). , 1093 ₂ , 1093 ₃ , 1093 ₄ are obtained.

FIG. 19 is a flowchart illustrating an example of processing when the processing for expanding correspondences between words according to the present embodiment is applied to machine translation (translation processing).
In FIG. 19, first, in step S901, a Chinese word corresponding to a Japanese word is obtained using a bilingual word dictionary from Japanese to Chinese (Japanese-Chinese bilingual dictionary) as one process during machine translation. Searched.

  In step S902, it is determined whether the search result word (in Chinese notation) in step S901 exists. If it is determined in step S902 that the corresponding word (in Chinese notation) exists, the series of processing ends.

  On the other hand, if it is determined in step S902 that the corresponding word (Chinese notation) does not exist, the process proceeds to step S903, and the search result is searched using a bilingual word dictionary from Japanese to English (Japanese-English bilingual dictionary). An English word corresponding to a Japanese word is searched.

  In step S904, the search result words (in English) in step S903 are stored in a list (one-dimensional array) L3 and sorted in the order of appearance frequency. In step S905, it is determined whether or not an English word (data) exists in the list L3.

  If it is determined in step S905 that a word (English notation) exists, the process proceeds to step S906, and the first word (English notation) is extracted from the list L3, stored in the buffer Temp1, and the word (English notation) stored in the Temp1. ) Is deleted from the list L3.

  In step S907, a Japanese word corresponding to the word (English notation) stored in Temp1 is searched using an English-to-Japanese bilingual word dictionary (English-Japanese bilingual dictionary). In step S908, The search result word (Japanese notation) in step S907 is added (stored) to the list L1. Then, the process returns to step S905.

  On the other hand, if it is determined in step S905 that the word (English notation) does not exist, the process proceeds to step S909, and it is determined whether a word (data) in Japanese notation exists in the list L1.

  If it is determined in step S909 that a word (Japanese notation) exists, the process proceeds to step S910, the first word (Japanese notation) is extracted from the list L1, stored in the buffer Temp2, and the word ( The word corresponding to (in Japanese) is deleted from the list L1.

  In step S911, using a bilingual word dictionary from Japanese to Chinese (Japanese-Chinese bilingual dictionary), a word in the second language (Chinese) corresponding to the word (Japanese notation) stored in Temp2 is obtained. A search is performed, and in step S912, it is determined whether or not there is a word (in Chinese notation) as a search result in step S911. If it is determined in step S912 that the search result word (in Chinese notation) exists, in step S913, the first word (in Chinese notation) in the search result list is the Japanese whose corresponding word is unknown. As a translation of the word.

On the other hand, if it is determined in step S912 that the search result word (in Chinese) does not exist, the process returns to step S909.
If it is determined in step S909 that no word (Japanese notation) exists, the series of processing ends.

  FIG. 19 shows the case where the processing by the second expansion processing execution unit 12 in FIG. 1A is applied to machine translation, but the other processing by the fourth or sixth expansion processing execution unit is machine translation. You may apply to.

  Further, the inter-word correspondence determination process of the present embodiment may be performed when the parallel translation example dictionary is registered, or may be performed when the parallel translation example is displayed to the user. That is, at the time of registering the bilingual example dictionary, the word division information constituting the example as a result of the morphological analysis and the correspondence information between words that are the processing result of the inter-word correspondence determining process of this embodiment are stored in the bilingual example dictionary You may link and store, and you may make it perform the morphological analysis and the correspondence determination process between words of this embodiment in the case of a display.

  The correspondence expansion processing unit between words according to the present embodiment can be configured as software. FIG. 20 is a diagram showing a hardware environment when each embodiment of the present invention is realized by a program.

  In FIG. 20, a computer as hardware includes a CPU 61, a memory 62, an input device 63, an output device 64, an external storage device 65, a medium drive device 66 capable of driving a portable recording medium 69, and a network connection device. 67 is connected via a bus 68.

  In FIG. 20, a CPU 61 controls the entire computer, and a memory 62 temporarily stores data stored in a secondary storage including the external storage device 65 when executing a program, updating data, or the like. Primary memory.

  The user can specify a sentence / translation pair in the bilingual example dictionary or issue a start instruction to a program that performs processing according to the present embodiment via the input device 63. Further, the processing result of the program that executes the processing of this embodiment can be viewed via the output device 64.

  In addition to the programs and data stored in the external storage device 65, the programs and data in the portable recording medium 69 are read through the medium and the network connection device 67, the programs and data read through the medium drive device 66, and so on. Information provider programs and data can be used inside the computer.

FIG. 21 is a diagram for explaining loading of a program.
The screen output data generation process of the present invention can naturally be realized by a general computer. In this case, the program for the processing of the present invention is loaded from the storage device 72 of the computer into the memory of the computer and executed, or the program for the processing of the present invention is stored in the memory of the computer from the portable recording medium 73. It is also possible to load and execute the program for the processing of the present invention from the storage device 71 side of the program provider 70 to the memory of the computer via the network.

The present invention may have the following configuration.
(Supplementary note 1) In a parallel translation example dictionary between a first language and a second language in which a correspondence relationship is set between a sentence in a first language and a translated sentence in a second language, the words constituting the sentence in the first language and the second In a program for causing a computer to execute a process of determining a correspondence relationship between words constituting a translation of a language using a bilingual word dictionary that creates a correspondence relationship between words in different languages,
Extracting a word in the third language corresponding to a word in the first language using a bilingual word dictionary from the first language to the third language;
Extracting a first language word corresponding to the extracted third language word using a bilingual word dictionary from the third language to the first language; and
Using a bilingual word dictionary from the first language to the second language, a word in the second language corresponding to the word in the first language in step 1 or the word in the first language extracted in step 2 Extracting as a candidate for translation;
When the translation candidate is included in a word constituting the translation of the second language, the translation candidate is determined as a translation corresponding to the word of the first language (in step 1);
A program for determining the correspondence between words in a bilingual example dictionary between a first language and a second language, wherein the computer executes:
(Supplementary Note 2) In a bilingual example dictionary between a first language and a second language in which a correspondence relationship is set between a sentence in the second language and a translated sentence in the first language, the words constituting the sentence in the second language and the first In a program for causing a computer to execute a process of determining a correspondence relationship between words constituting a translation of a language using a bilingual word dictionary that creates a correspondence relationship between words in different languages,
Extracting a word in the first language corresponding to a word in the second language as a first translation word candidate using a bilingual word dictionary from the second language to the first language;
Using a bilingual word dictionary from the first language to the third language to extract a third language word corresponding to the first translated word candidate of the extracted first language word;
Using the bilingual word dictionary from the third language to the first language to extract the first language word corresponding to the extracted third language word as a second translated word candidate;
When the first translation candidate or the second translation candidate is included in a word constituting the translation of the first language, the first translation candidate or the second translation candidate is selected as the first translation candidate extraction step. Determining) as a translation corresponding to a second language word;
A program for determining the correspondence between words in a bilingual example dictionary between a first language and a second language, wherein the computer executes:
(Supplementary Note 3) In a bilingual example dictionary between a first language and a second language in which a correspondence relationship is set between a sentence in a first language and a translated sentence in a second language, the words constituting the sentence in the first language and the second In a program for causing a computer to execute processing for determining a correspondence relationship between words constituting a translation of a language, using a parallel translation word dictionary that creates a correspondence relationship between words in different languages,
Extracting a semantic symbol corresponding to a word in the first language;
Extracting a word in the first language corresponding to the extracted semantic symbol;
Using the bilingual word dictionary from the first language to the second language, the second language corresponding to the first language word in step 1 or the first language word extracted in step 2 Extracting words as candidate translations;
When the translation candidate is included in a word constituting the translation of the second language, the translation candidate is determined as a translation corresponding to the word of the first language (in step 1);
A program for determining the correspondence between words in a bilingual example dictionary between a first language and a second language, wherein the computer executes:
(Supplementary Note 4) In a parallel translation example dictionary between the first language and the second language in which a correspondence relationship is set between the sentence in the second language and the translated sentence in the first language, the first word and the word constituting the sentence in the second language In a program for causing a computer to execute processing for determining a correspondence relationship between words constituting a translation of a language, using a parallel translation word dictionary that creates a correspondence relationship between words in different languages,
Extracting a word in the first language corresponding to a word in the second language as a first translation word candidate using a bilingual word dictionary from the second language to the first language;
Extracting a semantic symbol of the extracted first language word;
Extracting a word in the first language corresponding to the extracted semantic symbol as a second translation word candidate;
When the first translation candidate or the second translation candidate is included in a word constituting the translation of the first language, the first translation candidate or the second translation candidate is extracted as the first translation candidate extraction. Determining as a translation corresponding to a second language word of step);
A program for determining the correspondence between words in a bilingual example dictionary between a first language and a second language, wherein the computer executes:
(Supplementary Note 5) In a bilingual example dictionary between a first language and a second language in which a correspondence relationship is set between a sentence in the first language and a translated sentence in the second language, the words constituting the sentence in the second language and the first In a program for causing a computer to execute a process of determining a correspondence relationship between words constituting a translation of a language using a bilingual word dictionary that creates a correspondence relationship between words in different languages,
Using the bilingual word dictionary with meaning symbols in which correspondence between the words in the first language and the third language is given and the words in each language are also associated with meaning symbols, A first semantic symbol extraction step for extracting a semantic symbol;
Extracting a word in a third language corresponding to the extracted semantic symbol using the bilingual word dictionary with semantic symbols;
A second semantic symbol extraction step of extracting a semantic symbol corresponding to the extracted third language word using the bilingual word dictionary with semantic symbols;
Using the bilingual word dictionary with semantic symbols, the semantic language extracted corresponding to the first language word and the first language word corresponding to the semantic symbol extracted corresponding to the third language word Extracting the
Using a bilingual word dictionary from the first language to the second language, the second language corresponding to the first language word in the first semantic symbol extraction step and the extracted first language word Extracting words as candidate translations;
When the translation candidate is included in a word constituting the translation of the second language, the translation candidate is determined as a translation corresponding to the first language word in the first semantic symbol extraction step; ,
A program for determining the correspondence between words in a bilingual example dictionary between a first language and a second language, wherein the computer executes:
(Supplementary Note 6) In a parallel translation example dictionary between a first language and a second language in which a correspondence relationship is set between a sentence in a second language and a translated sentence in the first language, the words constituting the sentence in the second language and the first In a program for causing a computer to execute a process of determining a correspondence relationship between words constituting a translation of a language using a bilingual word dictionary that creates a correspondence relationship between words in different languages,
Extracting a word in the first language corresponding to a word in the second language as a first translation word candidate using a bilingual word dictionary from the second language to the first language;
Using the bilingual word dictionary with meaning symbols in which the words in the first language and the third language are associated with each other and the words in the respective languages are also associated with the meaning symbols, Extracting a semantic symbol of the word;
Extracting a word in a third language corresponding to the extracted semantic symbol using the bilingual word dictionary with semantic symbols;
Extracting a semantic symbol corresponding to the extracted word in the third language using the bilingual word dictionary with semantic symbols;
Extracting the semantic symbols extracted corresponding to the first language words and the first language words corresponding to the semantic symbols extracted corresponding to the third language words as second translation candidates;
In the case where the first translated word candidate or the second translated word candidate is included in a word constituting the sentence in the first language, the first translated word candidate or the second translated word candidate is extracted as the first translated word candidate. Determining as a translation corresponding to a second language word of
A program for determining the correspondence between words in a bilingual example dictionary between a first language and a second language, wherein the computer executes:
(Supplementary note 7) In the bilingual example dictionary between the first language and the second language in which a correspondence relationship is set between the example of the first language that is the translation of the other and the example of the second language, A process for determining a correspondence between a word constituting a sentence and a word constituting a translation in the second language, or a correspondence between a word constituting a sentence in the second language and a word constituting a translation in the first language; In a program to be executed by a computer using a bilingual word dictionary that provides correspondence between words in different languages,
A process related to the first, second, third, fourth, fifth, or sixth provisional decision step for temporarily deciding a translation word for the word in the first language example or the word in the second language example Can execute two or more processes specified in advance,
In the step relating to the first provisional decision,
Extracting a word in the first language corresponding to a word in the second language as a first translation word candidate using a bilingual word dictionary from the second language to the first language; Extracting a third language word corresponding to the extracted first language word using a bilingual word dictionary into three languages, and using a bilingual word dictionary from the third language to the first language, The step of extracting the first language word corresponding to the extracted third language word as a second translation word candidate, and the first translation word candidate or the second translation word candidate constitutes an example of the first language A first tentative determination step for tentatively determining the first translation candidate or the second translation candidate as a translation corresponding to the second language word in the first translation candidate extraction step, , Let the computer execute
In the step relating to the second provisional decision,
Extracting a word in the third language corresponding to a word in the first language using a bilingual word dictionary from the first language to the third language; and from the third language to the first language. Step 2 for extracting a first language word corresponding to the extracted third language word using the bilingual word dictionary of the first language, and using the bilingual word dictionary from the first language to the second language, Extracting a second language word corresponding to the first language word in step 1 or the first language word extracted in step 2 as a translation word candidate; and A second provisional decision step of provisionally determining the candidate word as a translation corresponding to the word in the first language in step 1 when the word is included in the word constituting the example ,
In the step related to the third provisional decision,
Extracting a word in the first language corresponding to a word in the second language as a first translated word candidate using a bilingual word dictionary from the second language to the first language, and the extracted first language Extracting a semantic symbol of the first word, extracting a word in the first language corresponding to the extracted semantic symbol as a second translation word candidate, the first translation word candidate or the second translation word candidate being the first translation word A first translation candidate or a second translation candidate is temporarily determined as a translation corresponding to a second language word in the first translation candidate extraction step. 3 provisional determination steps are executed by the computer,
In the step related to the fourth provisional decision,
Step 1 for extracting semantic symbols corresponding to the words in the first language, Step 2 for extracting words in the first language corresponding to the extracted semantic symbols, and conversion from the first language to the second language Extracting a second language word corresponding to the first language word in step 1 or the first language word extracted in step 2 as a translation word candidate using a bilingual word dictionary; A fourth tentative determination step of tentatively determining a translation word candidate as a translation corresponding to the word of the first language in step 1 when the translation word candidate is included in a word constituting the example of the second language And causing the computer to execute
In the step related to the fifth provisional decision,
Extracting a word in the first language corresponding to a word in the second language as a first translation word candidate using a bilingual word dictionary from the second language to the first language; Extract semantic symbols of extracted words in the first language using a bilingual word dictionary with semantic symbols that associates the words with the three languages and associates the words of each language with semantic symbols. Extracting a word in a third language corresponding to the extracted semantic symbol using the bilingual word dictionary with semantic symbols, and extracting a third word extracted using the bilingual word dictionary with semantic symbols. Extracting a semantic symbol corresponding to a language word; a semantic symbol extracted corresponding to a first language word; and a first language corresponding to a semantic symbol extracted corresponding to a third language word. Candidate second translation word Extracting the first translation candidate or the second translation candidate if the first translation candidate or the second translation candidate is included in a word constituting the sentence of the first language. , Causing the computer to execute a fifth tentative determination step of tentatively determining a translation corresponding to a word in the second language in the first translation candidate extraction step
In the step related to the sixth provisional decision,
Using the bilingual word dictionary with meaning symbols in which correspondence between the words in the first language and the third language is given and the words in each language are also associated with meaning symbols, A first semantic symbol extracting step of extracting a semantic symbol; a step of extracting a third language word corresponding to the extracted semantic symbol using the semantic symbol-added bilingual word dictionary; and the semantic symbol-added bilingual word dictionary A second semantic symbol extracting step for extracting a semantic symbol corresponding to the extracted word in the third language, and an extraction corresponding to the word in the first language using the bilingual word dictionary with semantic symbols Extracting a first language word corresponding to the extracted semantic symbol and the semantic language extracted corresponding to the third language word, and using a bilingual word dictionary from the first language to the second language The first Extracting a second language word corresponding to the first language word in the semantic symbol extraction step and the extracted first language word as a translation word candidate, and the translation word candidate is an example of the second language. A sixth provisional decision step for provisionally deciding a candidate word as a translation corresponding to a word of a sentence in a first language in the first semantic symbol extraction step when it is included in a constituent word; To run
Furthermore, when there are a plurality of translated words when the result of the provisional decision of the provisional decision step corresponding to the process designated to be executed is reflected, the process having priority between the two or more executed processes The result is determined so that the result is reflected in the translated word, and when there are not a plurality of translated words, the computer is caused to execute a step of determining the combined result of the tentative determination step as a translated word as it is A program for determining correspondence between words in a parallel translation example dictionary between a first language and a second language.
(Supplementary Note 8) In a bilingual example dictionary between a first language and a second language in which a correspondence relationship is set between a sentence in a first language and a translated sentence in a second language, the words constituting the sentence in the first language and the second In a method in which a computer executes a process of determining a correspondence relationship between words constituting a translation of a language using a parallel translation word dictionary that creates a correspondence relationship between words in different languages,
Extracting a word in the third language corresponding to a word in the first language using a bilingual word dictionary from the first language to the third language;
Extracting a first language word corresponding to the extracted third language word using a bilingual word dictionary from the third language to the first language; and
Using the bilingual word dictionary from the first language to the second language, the second language corresponding to the first language word in step 1 or the first language word extracted in step 2 Extracting words as candidate translations;
When the translation candidate is included in a word constituting the translation of the second language, the translation candidate is determined as a translation corresponding to the word of the first language (in step 1);
A method for determining a correspondence relationship between words of a parallel translation example dictionary between a first language and a second language.
(Supplementary Note 9) In a bilingual example dictionary between a first language and a second language in which a correspondence relationship is set between a sentence in the second language and a translated sentence in the first language, the words constituting the sentence in the second language and the first In a method in which a computer executes a process of determining a correspondence relationship between words constituting a translation of a language using a parallel translation word dictionary that creates a correspondence relationship between words in different languages,
Extracting a word in the first language corresponding to a word in the second language as a first translation word candidate using a bilingual word dictionary from the second language to the first language;
Using a bilingual word dictionary from the first language to the third language to extract a third language word corresponding to the first translated word candidate of the extracted first language word;
Extracting the first language word corresponding to the extracted third language word as a second translation word candidate;
When the first translation candidate or the second translation candidate is included in a word constituting the sentence in the first language, the first translation candidate or the second translation candidate is extracted as the first translation candidate extraction. Determining as a translation corresponding to a second language word of step);
A method for determining a correspondence relationship between words of a parallel translation example dictionary between a first language and a second language.
(Supplementary Note 10) In a parallel translation example dictionary between a first language and a second language in which a correspondence relationship is set between a sentence in a first language and a translated sentence in a second language, the words constituting the sentence in the second language and the first In a method in which a computer executes a process of determining a correspondence relationship between words constituting a translation of a language using a parallel translation word dictionary that creates a correspondence relationship between words in different languages,
Extracting a semantic symbol corresponding to a word constituting the sentence of the first language;
Extracting another word in the first language corresponding to the extracted semantic symbol;
Using the bilingual word dictionary from the first language to the second language, the second language corresponding to the first language word in step 1 or the first language word extracted in step 2 Extracting words as candidate translations;
When the translation candidate is included in a word constituting the translation of the second language, the translation candidate is determined as a translation corresponding to the word of the first language (in step 1);
A method for determining a correspondence relationship between words of a parallel translation example dictionary between a first language and a second language.
(Supplementary Note 11) In a bilingual example dictionary between a first language and a second language in which a correspondence relationship is set between a sentence in a second language and a translated sentence in the first language, the words constituting the sentence in the second language and the first In a method in which a computer executes a process of determining a correspondence relationship between words constituting a translation of a language using a parallel translation word dictionary that creates a correspondence relationship between words in different languages,
Extracting a word in the first language corresponding to a word in the second language as a first translation word candidate using a bilingual word dictionary from the second language to the first language;
Extracting a semantic symbol of the extracted first language word;
Extracting a word in the first language corresponding to the extracted semantic symbol as a second translation word candidate;
When the first translation candidate or the second translation candidate is included in a word constituting the sentence in the first language, the first translation candidate or the second translation candidate is extracted as the first translation candidate extraction. Determining as a translation corresponding to a second language word of step);
A method for determining a correspondence relationship between words of a parallel translation example dictionary between a first language and a second language.
(Supplementary Note 12) In a bilingual example dictionary between a first language and a second language in which a correspondence relationship is set between a sentence in a first language and a translated sentence in a second language, the words constituting the sentence in the second language and the first In a method in which a computer executes a process of determining a correspondence relationship between words constituting a translation of a language using a parallel translation word dictionary that creates a correspondence relationship between words in different languages,
A word in the first language is created using a bilingual word dictionary with meaning symbols in which correspondence between the words in the first language and the third language is provided and the words in the respective languages are also associated with meaning symbols. Performing a first semantic symbol extraction to extract a semantic symbol of
Extracting a word in a third language corresponding to the extracted semantic symbol using the bilingual word dictionary with semantic symbols;
Performing second semantic symbol extraction using the bilingual word dictionary with semantic symbols to extract semantic symbols corresponding to the extracted third language words;
Using the bilingual word dictionary with semantic symbols, the semantic language extracted corresponding to the first language word and the first language word corresponding to the semantic symbol extracted corresponding to the third language word Extracting the
Using a bilingual word dictionary from the first language to the second language, a second language word corresponding to the first language word in the first semantic symbol extraction and the extracted first language word Extracting as a candidate for translation;
When the translation candidate is included in a word constituting the translation in the second language, the translation candidate is determined as a translation corresponding to the first language word in the first semantic symbol extraction;
A method for determining a correspondence relationship between words of a parallel translation example dictionary between a first language and a second language.
(Supplementary note 13) In a parallel translation example dictionary between the first language and the second language in which a correspondence relationship is set between the sentence in the second language and the translation in the first language, the first word and the word constituting the sentence in the second language In a method in which a computer executes a process of determining a correspondence relationship between words constituting a translation of a language using a parallel translation word dictionary that creates a correspondence relationship between words in different languages,
Extracting a word in the first language corresponding to a word in the second language as a first translation word candidate using a bilingual word dictionary from the second language to the first language;
Using the bilingual word dictionary with meaning symbols in which the words in the first language and the third language are associated with each other and the words in the respective languages are also associated with the meaning symbols, Extracting a semantic symbol of the word;
Extracting a word in a third language corresponding to the extracted semantic symbol using the bilingual word dictionary with semantic symbols;
Extracting a semantic symbol corresponding to the extracted word in the third language using the bilingual word dictionary with semantic symbols;
Extracting the semantic symbols extracted corresponding to the first language words and the first language words corresponding to the semantic symbols extracted corresponding to the third language words as second translation candidates;
In the case where the first translated word candidate or the second translated word candidate is included in a word constituting the sentence in the first language, the first translated word candidate or the second translated word candidate is extracted as the first translated word candidate. Determining as a translation corresponding to a second language word of
A method for determining a correspondence relationship between words of a parallel translation example dictionary between a first language and a second language.
(Supplementary note 14) In the bilingual example dictionary between the first language and the second language in which a correspondence relationship is set between the example of the first language and the example of the second language, each of which is the other translation, the first language A process for determining a correspondence between a word constituting a sentence and a word constituting a translation in the second language, or a correspondence between a word constituting a sentence in the second language and a word constituting a translation in the first language; In a method executed by a computer using a bilingual word dictionary that establishes correspondence between words in different languages,
A process related to the first, second, third, fourth, fifth, or sixth provisional decision that temporarily decides a translation for the word in the first language example or the word in the second language example. Among them, two or more processes specified in advance can be executed,
The process related to the first provisional decision is as follows.
Extracting a word in the first language corresponding to a word in the second language as a first translation word candidate using a bilingual word dictionary from the second language to the first language; Extracting a third language word corresponding to the extracted first language word using a bilingual word dictionary into three languages, and using a bilingual word dictionary from the third language to the first language, The step of extracting the first language word corresponding to the extracted third language word as a second translation word candidate, and the first translation word candidate or the second translation word candidate constitutes an example of the first language The first translation candidate or the second translation candidate is tentatively determined as a translation corresponding to the second language word of the first translation candidate extraction step,
The process related to the second provisional decision is as follows:
Extracting a word in the third language corresponding to a word in the first language using a bilingual word dictionary from the first language to the third language; and from the third language to the first language. Step 2 for extracting a first language word corresponding to the extracted third language word using the bilingual word dictionary of the first language, and using the bilingual word dictionary from the first language to the second language, Extracting a second language word corresponding to the first language word in step 1 or the first language word extracted in step 2 as a translation word candidate; and A step of tentatively determining the translation candidate as a translation corresponding to the word in the first language in step 1,
The process related to the third provisional decision is as follows.
Extracting a word in the first language corresponding to a word in the second language as a first translated word candidate using a bilingual word dictionary from the second language to the first language, and the extracted first language Extracting a semantic symbol of the first word, extracting a word in the first language corresponding to the extracted semantic symbol as a second translation word candidate, the first translation word candidate or the second translation word candidate being the first translation word A step of tentatively determining the first translation candidate or the second translation candidate as a translation corresponding to the second language word in the first translation candidate extraction step, if it is included in the word constituting the language example And
The process related to the fourth provisional decision is as follows.
Step 1 for extracting semantic symbols corresponding to the words in the first language, Step 2 for extracting words in the first language corresponding to the extracted semantic symbols, and conversion from the first language to the second language Extracting a second language word corresponding to the first language word in step 1 or the first language word extracted in step 2 as a translation word candidate using a bilingual word dictionary; When a translation candidate is included in a word constituting an example of the second language, the translation candidate is provisionally determined as a translation corresponding to the word in the first language in step 1;
The process related to the fifth provisional decision is as follows.
Extracting a word in the first language corresponding to a word in the second language as a first translation word candidate using a bilingual word dictionary from the second language to the first language; Extract semantic symbols of extracted words in the first language using a bilingual word dictionary with semantic symbols that associates the words with the three languages and associates the words of each language with semantic symbols. Extracting a word in a third language corresponding to the extracted semantic symbol using the bilingual word dictionary with semantic symbols, and extracting a third word extracted using the bilingual word dictionary with semantic symbols. Extracting a semantic symbol corresponding to a language word; a semantic symbol extracted corresponding to a first language word; and a first language corresponding to a semantic symbol extracted corresponding to a third language word. Candidate second translation word Extracting the first translation candidate or the second translation candidate if the first translation candidate or the second translation candidate is included in a word constituting the sentence of the first language. Tentatively determining as a translation corresponding to a word in the second language in the first translation candidate extraction,
The process related to the sixth provisional decision is as follows.
Using the bilingual word dictionary with meaning symbols in which correspondence between the words in the first language and the third language is given and the words in each language are also associated with meaning symbols, Performing a first semantic symbol extraction for extracting a semantic symbol; extracting a third language word corresponding to the extracted semantic symbol using the bilingual word dictionary with semantic symbols; and the semantic symbol Executing a second semantic symbol extraction for extracting a semantic symbol corresponding to the extracted word in the third language using the bilingual translated word dictionary; and using the bilingual word dictionary with the semantic symbol for the first Extracting a semantic symbol extracted corresponding to a language word and a first language word corresponding to a semantic symbol extracted corresponding to the third language word; and from the first language to the second language Translation into language Extracting a second language word corresponding to the first language word in the first semantic symbol extraction and the extracted first language word as a translation word candidate using a word dictionary; and Provisionally determining a translation word candidate as a translation word corresponding to a word of a sentence in the first language in the first semantic symbol extraction when it is included in a word constituting an example of the second language. ,
Furthermore, when there are a plurality of translated words when the result of the provisional decision of the provisional decision step corresponding to the process designated to be executed is reflected, the process having priority between the two or more executed processes The first language and the second language are characterized in that the result is determined so as to be reflected in the translated word, and when a plurality of translated words do not exist, the combined result of the tentative determination in the tentative determination step is determined as a translated word as it is. A method for determining correspondence between words in an example dictionary for translation between languages.
(Additional remark 15) In the translation processing program which makes a computer perform the process which translates the sentence of a 1st language into a 2nd language using the bilingual word dictionary from a 1st language to a 2nd language,
If the second language word corresponding to the word constituting the sentence in the first language does not exist even if the bilingual word dictionary from the first language to the second language is used, the first language to the third language Extracting a word in the third language corresponding to a word in the first language using a bilingual word dictionary
Extracting a first language word corresponding to the extracted third language word using a bilingual word dictionary from the third language to the first language;
Extracting a second language word corresponding to the first language word extracted in the first language word extraction step as a translation word candidate using a bilingual word dictionary from the first language to the second language. When,
When any of the translation candidates exists, determining any of the translation candidates as a translation;
A translation processing program for causing a computer to execute
(Additional remark 16) In the translation processing program which makes a computer perform the process which translates the sentence of a 1st language into a 2nd language using the bilingual word dictionary from a 1st language to a 2nd language,
If a second language word corresponding to a word constituting the sentence of the first language does not exist even if a bilingual word dictionary from the first language to the second language is used, the word of the first language Extracting corresponding semantic symbols;
Extracting a first language word corresponding to the extracted semantic symbols;
Extracting a second language word corresponding to the first language word extracted in the first language word extraction step as a translation word candidate using a bilingual word dictionary from the first language to the second language. When,
When any of the translation candidates exists, determining any of the translation candidates as a translation;
A translation processing program for causing a computer to execute
(Additional remark 17) In the translation processing program which makes a computer perform the process which translates the sentence of a 1st language into a 2nd language using the bilingual word dictionary from a 1st language to a 2nd language,
When the second language word corresponding to the word constituting the sentence of the first language does not exist even when the bilingual word dictionary from the first language to the second language is used, the first language and the third language Extracting a semantic symbol of a word in the first language using a bilingual word dictionary with semantic symbols in which the words in each language are also associated with semantic symbols; ,
Extracting a word in a third language corresponding to the extracted semantic symbol using the bilingual word dictionary with semantic symbols;
Extracting a semantic symbol corresponding to the extracted word in the third language using the bilingual word dictionary with semantic symbols;
Using the bilingual word dictionary with semantic symbols, the semantic language extracted corresponding to the first language word and the first language word corresponding to the semantic symbol extracted corresponding to the third language word Extracting the
Extracting a second language word corresponding to the extracted first language word as a translation word candidate using a bilingual word dictionary from the first language to the second language;
When any of the translation candidates exists, determining any of the translation candidates as a translation;
A translation processing program for causing a computer to execute
(Supplementary Note 18) In a translation processing method in which a computer executes a process of translating a sentence in a first language into a second language using a bilingual word dictionary from the first language to the second language.
If the second language word corresponding to the word constituting the sentence in the first language does not exist even if the bilingual word dictionary from the first language to the second language is used, the first language to the third language Extracting a word in the third language corresponding to a word in the first language using a bilingual word dictionary
Using a bilingual word dictionary from the third language to the first language, extracting words in the first language corresponding to the extracted words in the third language;
Using a bilingual word dictionary from the first language to the second language, extracting a second language word corresponding to the first language word extracted in the first language word extraction step as a translation word candidate;
If there is a translation candidate, determine one of the translation candidates as a translation;
A translation processing method characterized by the above.
(Additional remark 19) In the translation processing method which a computer performs the process which translates the sentence of a 1st language into a 2nd language using the bilingual word dictionary from a 1st language to a 2nd language,
If a second language word corresponding to a word constituting the sentence of the first language does not exist even if a bilingual word dictionary from the first language to the second language is used, the word of the first language Extract the corresponding semantic symbols,
Extracting a first language word corresponding to the extracted semantic symbol;
Using a bilingual word dictionary from the first language to the second language, extracting a second language word corresponding to the first language word extracted in the first language word extraction step as a translation word candidate;
If there is a translation candidate, determine one of the translation candidates as a translation;
A translation processing method characterized by the above.
(Additional remark 20) In the translation processing method which a computer performs the process which translates the sentence of a 1st language into a 2nd language using the bilingual word dictionary from a 1st language to a 2nd language,
When the second language word corresponding to the word constituting the sentence of the first language does not exist even when the bilingual word dictionary from the first language to the second language is used, the first language and the third language And using the bilingual word dictionary with meaning symbols in which the words in the respective languages are also associated with the meaning symbols, the meaning symbols of the words in the first language are extracted,
Extracting a word in a third language corresponding to the extracted semantic symbol using the bilingual word dictionary with semantic symbols;
Using the bilingual word dictionary with semantic symbols, extracting semantic symbols corresponding to the extracted words in the third language;
Using the bilingual word dictionary with semantic symbols, the semantic language extracted corresponding to the first language word and the first language word corresponding to the semantic symbol extracted corresponding to the third language word Extract
Using a bilingual word dictionary from the first language to the second language, extracting a second language word corresponding to the extracted first language word as a translation word candidate;
If there is a translation candidate, determine one of the translation candidates as a translation;
A translation processing method characterized by the above.

It is a block diagram which shows the structure of the correspondence determination processing part between words of the example dictionary of a Japanese-Chinese translation according to one Embodiment of this invention. It is the modification of the structure of the unit in connection with the write-in process of the corresponding information of FIG. 1A. It is a figure explaining operation | movement of the correspondence relationship determination process part between words of this embodiment. It is a flowchart which shows an example of the determination process of the correspondence between words of this embodiment. 3 is a correspondence matrix corresponding to the Japanese example sentence and the Chinese example sentence of FIG. FIG. 4 is a flowchart showing in more detail the writing process to the correspondence matrix executed in FIG. 3. FIG. It is a figure explaining the process (the 1) which expands the range of the Japanese word (candidate) corresponding to a Chinese word. It is a flowchart of the process (the 1) which expands the range of the Japanese word (candidate) corresponding to a Chinese word. It is a figure explaining the process (the 1) which expands the range of the Chinese word (candidate) corresponding to a Japanese word. It is a flowchart of the process (the 1) which expands the range of the Chinese word (candidate) corresponding to a Japanese word. It is a figure explaining the process (the 2) which expands the range of the Japanese word (candidate) corresponding to a Chinese word. It is a flowchart of the process (the 2) which expands the range of the Japanese word (candidate) corresponding to a Chinese word. It is a figure explaining the process (the 2) which expands the range of the Chinese word (candidate) corresponding to a Japanese word. It is a flowchart of the process (the 2) which expands the range of the Chinese word (candidate) corresponding to a Japanese word. It is a figure explaining the process (the 3) which expands the range of the Japanese word (candidate) corresponding to a Chinese word. It is a flowchart of the process (the 3) which expands the range of the Japanese word (candidate) corresponding to a Chinese word. It is a figure explaining the process (the 3) which expands the range of the Chinese word (candidate) corresponding to a Japanese word. It is a flowchart of the process (the 3) which expands the range of the Chinese word (candidate) corresponding to a Japanese word. It is a figure explaining the case where the expansion processing of the correspondence between words of this embodiment is applied to machine translation (translation processing). It is a flowchart which shows an example of a process at the time of applying the expansion process of the correspondence between the words of this embodiment to machine translation (translation processing). It is a figure which shows the hardware environment in the case of implement | achieving each embodiment of this invention with a program. It is a figure explaining loading of a program.

Explanation of symbols

DESCRIPTION OF SYMBOLS 10 Correspondence expansion processing part 11 1st expansion process execution part 12 2nd expansion process execution part 13 3rd expansion process execution part 14 4th expansion process execution part 15 5th expansion process execution part 16 6th expansion process execution part 17 Search Part (extraction part)
18 control unit 19 hit determination unit 21 first corresponding word tentative determination unit 22 second corresponding word tentative determination unit 23 third corresponding word tentative determination unit 24 fourth corresponding word tentative determination unit 25 fifth corresponding word tentative determination unit 26 sixth Corresponding word provisional determination unit 27 Uniqueness determination unit

Claims (8)

In a bilingual example dictionary between a first language and a second language in which a correspondence relationship is set between a sentence in a first language and a translation in a second language, the words constituting the sentence in the first language and the translation in the second language In a program for causing a computer to execute a process of determining a correspondence relationship between constituent translation words using a bilingual word dictionary that creates a correspondence relationship between words in different languages,
Extracting a word in the third language corresponding to a word in the first language using a bilingual word dictionary from the first language to the third language;
Extracting a first language word corresponding to the extracted third language word using a bilingual word dictionary from the third language to the first language; and
Using the bilingual word dictionary from the first language to the second language, the second language corresponding to the first language word in step 1 or the first language word extracted in step 2 Extracting words as candidate translations;
When the translation candidate is included in a word constituting the translation of the second language, the translation candidate is determined as a translation corresponding to the word of the first language;
A program for determining the correspondence between words in a bilingual example dictionary between a first language and a second language, wherein the computer executes: In a parallel translation example dictionary between a first language and a second language in which a correspondence relationship is set between a sentence in the second language and a translated sentence in the first language, the words constituting the sentence in the second language and the translation in the first language In a program for causing a computer to execute a process of determining a correspondence relationship between constituent translation words using a bilingual word dictionary that creates a correspondence relationship between words in different languages,
Extracting a word in the first language corresponding to a word in the second language as a first translation word candidate using a bilingual word dictionary from the second language to the first language;
Using the bilingual word dictionary from the first language to the third language to extract words in the third language corresponding to the extracted first translated word candidates;
Using the bilingual word dictionary from the third language to the first language to extract the first language word corresponding to the extracted third language word as a second translated word candidate;
When the first translated word candidate or the second translated word candidate is included in a word constituting the translated text of the first language, the first translated word candidate or the second translated word candidate corresponds to the word of the second language A step of determining the translated word to be
A program for determining the correspondence between words in a bilingual example dictionary between a first language and a second language, wherein the computer executes: In a bilingual example dictionary between a first language and a second language in which a correspondence relationship is set between a sentence in a first language and a translation in a second language, the words constituting the sentence in the first language and the translation in the second language In a program for causing a computer to execute a process for determining a correspondence relationship with a translated word using a parallel word dictionary that creates a correspondence relationship between words in different languages,
Extracting a semantic symbol corresponding to a word in the first language;
Extracting a word in the first language corresponding to the extracted semantic symbol;
Using a bilingual word dictionary from the first language to the second language, a word in the second language corresponding to the word in the first language in step 1 or the word in the first language extracted in step 2 Extracting as a candidate for translation;
When the translation candidate is included in a word constituting the translation of the second language, the translation candidate is determined as a translation corresponding to the word of the first language;
A program for determining the correspondence between words in a bilingual example dictionary between a first language and a second language, wherein the computer executes: In a parallel translation example dictionary between a first language and a second language in which a correspondence relationship is set between a sentence in the second language and a translated sentence in the first language, the words constituting the sentence in the second language and the translation in the first language In a program for causing a computer to execute a process for determining a correspondence relationship with a translated word using a parallel word dictionary that creates a correspondence relationship between words in different languages,
Extracting a word in the first language corresponding to a word in the second language as a first translation word candidate using a bilingual word dictionary from the second language to the first language;
Extracting a semantic symbol of the extracted first language word;
Extracting a word in the first language corresponding to the extracted semantic symbol as a second translation word candidate;
When the first translation word candidate or the second translation word candidate is included in a word constituting the translation in the first language, the first translation word candidate or the second translation word candidate is converted into a word in the second language. Determining a corresponding translation,
A program for determining the correspondence between words in a bilingual example dictionary between a first language and a second language, wherein the computer executes: In a bilingual example dictionary between a first language and a second language in which a correspondence relationship is set between a sentence in a first language and a translation in a second language, the words constituting the sentence in the first language and the translation in the second language In a method in which a computer executes a process of determining a correspondence relationship between constituent translation words using a bilingual word dictionary that creates a correspondence relationship between words in different languages.
Extracting a word in the third language corresponding to a word in the first language using a bilingual word dictionary from the first language to the third language;
Extracting a first language word corresponding to the extracted third language word using a bilingual word dictionary from the third language to the first language; and
Using a bilingual word dictionary from the first language to the second language, a word in the second language corresponding to the word in the first language in step 1 or the word in the first language extracted in step 2 Extracting as a candidate for translation;
When the translation candidate is included in a word constituting the translation of the second language, the translation candidate is determined as a translation corresponding to the word of the first language;
A method for determining a correspondence relationship between words of a parallel translation example dictionary between a first language and a second language. In a parallel translation example dictionary between a first language and a second language in which a correspondence relationship is set between a sentence in the second language and a translated sentence in the first language, the words constituting the sentence in the second language and the translation in the first language In a method in which a computer executes a process of determining a correspondence relationship between constituent translation words using a bilingual word dictionary that creates a correspondence relationship between words in different languages.
Extracting a word in the first language corresponding to a word in the second language as a first translation word candidate using a bilingual word dictionary from the second language to the first language;
Using the bilingual word dictionary from the first language to the third language to extract words in the third language corresponding to the extracted first translated word candidates;
Extracting the first language word corresponding to the extracted third language word as a second translation word candidate;
When the first translation candidate or the second translation candidate is included in a word constituting the sentence in the first language, the first translation candidate or the second translation candidate is used as a word in the second language. Determining a corresponding translation,
A method for determining a correspondence relationship between words of a parallel translation example dictionary between a first language and a second language. In a bilingual example dictionary between a first language and a second language in which a correspondence relationship is set between a sentence in a first language and a translation in a second language, the words constituting the sentence in the second language and the translation in the first language In a method in which a computer executes a process of determining a correspondence relationship between constituent translation words using a bilingual word dictionary that creates a correspondence relationship between words in different languages.
Extracting a semantic symbol corresponding to a word constituting the sentence of the first language;
Extracting a word in the first language corresponding to the extracted semantic symbol;
Using the bilingual word dictionary from the first language to the second language, the second language corresponding to the first language word in step 1 or the first language word extracted in step 2 Extracting words as candidate translations;
When the translation candidate is included in a word constituting the translation of the second language, the translation candidate is determined as a translation corresponding to the word of the first language;
A method for determining a correspondence relationship between words of a parallel translation example dictionary between a first language and a second language. In a parallel translation example dictionary between a first language and a second language in which a correspondence relationship is set between a sentence in the second language and a translated sentence in the first language, the words constituting the sentence in the second language and the translation in the first language In a method in which a computer executes a process of determining a correspondence relationship between constituent translation words using a bilingual word dictionary that creates a correspondence relationship between words in different languages.
Extracting a word in the first language corresponding to a word in the second language as a first translation word candidate using a bilingual word dictionary from the second language to the first language;
Extracting a semantic symbol of the extracted first language word;
Extracting a word in the first language corresponding to the extracted semantic symbol as a second translation word candidate;
When the first translation candidate or the second translation candidate is included in a word constituting the sentence in the first language, the first translation candidate or the second translation candidate is used as a word in the second language. Determining a corresponding translation,
A method for determining a correspondence relationship between words of a parallel translation example dictionary between a first language and a second language.

Images