JPH04299759A - Foreign language/square form of kana description converting system - Google Patents

Abstract

PURPOSE:To support the preparedness of various descriptions in the work of preparing the indexes of a dictionary and a data base by outputting the oscillation of the description generated when converting the words of the foreign language to the description in the square form of Japanese syllabary (KANA). CONSTITUTION:When the word of the foreign language written in phonograms is inputted from a word description input part 1, the pronounciation of the word is estimated by a foreign language analysis part 3 while using knowledge stored in a foreign language pronounciation analysis knowledge base 2. A pronounciation converting into Japanese part 5 converts the pronounciation of the source word into Japanese by using a pronounciation converting into Japanese knowledge base 4. In that case, when there is possibility of plural pronounciations for the Japanese pronounciation, all the pronounciations are outputted. The Japanese pronounciation is converted to the square form of KANA by a pronounciation square form of KANA description part 7 while using a pronounciation square form of KANA description knowledge base 6. In that case, when there is the possibility of plural descriptions for displaying the square form of KANA, all the descriptions are outputted and displayed at a Japanese description display part 8.

Description

[Detailed description of the invention]

0001

[Industrial Application Field] The present invention relates to a foreign language katakana notation conversion system, and in particular to Japanese word processors, machine translation,
This paper relates to a foreign language katakana notation conversion method when writing foreign languages in katakana in Japanese documents, such as databases with Japanese headings.

0002

BACKGROUND OF THE INVENTION In recent years, there has been a growing tendency for foreign language technical terms to be used in katakana in various specialized fields.

[0003] In the conventional Japanese katakana notation conversion method, when registering katakana technical terms etc. in dictionaries used in Japanese word processors, machine translations, etc., or in databases with Japanese headings, the registration worker converts the katakana notation. I was thinking about each one as I typed them. For example, 2 in katakana notation
Even when there was a standard notation conversion (writing style), one system relied on the registration worker to think about it.

0004

[Problem to be solved by the invention] However, in the conventional foreign language katakana notation conversion method, there are many variations in the notation in Japanese, and especially when a foreign language is written in katakana as Japanese, the conversion of the notation is a registration process. There is a problem that it may differ depending on the person.

For example, when translating the English word "interface" into Japanese, the first half of the word "inter" can be written in two ways: "inter" and "inter", and the second half of the word "face" can be written in at least two ways: There are four types: Face, Face, Face, and Face. As a result, the Japanese katakana notation for "interface" is "interface" and "
There are eight types such as "interface", "interface", and "interface".

[0006] In addition, in the conventional foreign language katakana notation conversion method, there is a variation in the notation as described above for words that should be treated as the same word, so the notation registered in the dictionary or database heading and the actual search This raises the problem that there may be a discrepancy between the notation and the notation entered at the time of the search, resulting in a search failure. For example, even if you enter "interface" in a situation where only "interface" is registered in a dictionary or database, you will not be able to search that dictionary or database.

[0007] In order to avoid these problems, it is necessary to devise all possible variants in some way. One method is to register all different spellings when creating a dictionary or database. Another method is to take into account different spellings when searching for them.

However, in the conventional method described above, it is extremely difficult to manually come up with such different notations. This is because, as you can see in the example of "interface" and "interface" above, Japanese people hardly notice the difference in the notation,
This is because coming up with all possible alternative notations is a very difficult task.

An object of the present invention is to provide support for automatically outputting the possibility of variations in spelling when writing foreign languages in katakana, and alerting dictionary or database creators to different spellings. The objective is to provide a foreign language katakana notation conversion method as an environment.

[0010]Furthermore, the present invention aims to reduce the number of man-hours required for registering different spellings, improve the quality of registering different spellings, and, as a result, reduce situations in which access to a dictionary or database becomes impossible due to fluctuations in spellings. do.

[0011]

[Means for Solving the Problems] The foreign language katakana notation conversion method of the present invention includes a word list input means for inputting foreign words written in phonetic characters, and a word list input means for inputting foreign words written in phonetic characters. A foreign language pronunciation analysis knowledge base that stores knowledge for deriving the pronunciation of the input word from its notation, and a foreign language pronunciation analysis means that searches this foreign language pronunciation analysis knowledge base to derive the pronunciation of the input word. , a pronunciation Japaneseization knowledge base that stores knowledge for estimating the pronunciation of a foreign word when it is translated into Japanese from the pronunciation and transcription of the foreign word, and this pronunciation Japaneseization knowledge base. a pronunciation Japanese pronunciation means for guiding the pronunciation when the inputted word is converted into Japanese; and a pronunciation katakana notation knowledge base storing knowledge for displaying the pronunciation obtained by the pronunciation Japanese pronunciation means in katakana. , a pronunciation katakana notation means that searches this pronunciation katakana notation knowledge base and derives the notation when the input word is written in katakana, and a word of the input foreign language based on the output of this pronunciation katakana notation means. and Japanese notation display means for outputting a plurality of possible Japanese notations corresponding to the Japanese notation.

0012

[Operation] The foreign language katakana notation conversion method of the present invention maintains the correspondence from foreign language pronunciation to Japanese katakana notation as a knowledge base, so the notation that occurs when a foreign language word is written in katakana is retained as a knowledge base. can output vibrations.

[0013] When words in a foreign language are written in katakana, variations in the spelling first occur at the stage of Japanese pronunciation. For example, the pronunciation of the English word "face" is [feis], but the Japanese pronunciation of the vowel [ei] is [
There are two possibilities: [ei] (``ei'') and [e:] (``ei''). In the present invention, since such knowledge is stored in the pronunciation Japanese conversion means, it is possible to output the fluctuations in spelling that occur at this stage.

[0014] In addition, fluctuations in writing can also occur when writing sounds that do not previously exist in Japanese in katakana. For example, the beginning [fe] of the Japanese pronunciation [feisu] of the English pronunciation [feis] of "face" does not exist in the original Japanese, so its katakana notation is not consistent. Specifically, there are two ways to write the [fe] part: "fe" and "hue". In the present invention, since such knowledge is stored in the pronunciation katakana notation means, it is possible to output the fluctuations in notation that occur at this stage.

[0015]

DESCRIPTION OF THE PREFERRED EMBODIMENTS Next, the present invention will be explained with reference to the drawings.

FIG. 1 is a block diagram showing a foreign language Katakana notation conversion support device according to an embodiment of the present invention.

In FIG. 1, the foreign language katakana orthography conversion support device of this embodiment includes a word table input section 1 into which words of a foreign language written in phonetic characters are input, and a foreign word input section 1 inputted from the word table input section 1. A foreign language pronunciation analysis knowledge base 2 that stores knowledge for deriving the pronunciation from the notation of the word, and a foreign language pronunciation that searches this foreign language pronunciation analysis knowledge base 2 to derive the pronunciation of the input foreign language word. an analysis unit 3, a pronunciation Japaneseization knowledge base 4 that stores knowledge for estimating the pronunciation of a word when it is translated into Japanese from the pronunciation and transcription of a foreign language word, and this pronunciation Japaneseization knowledge base 4. A pronunciation Japanese translation unit 5 that searches for and guides the pronunciation when the word is translated into Japanese, and pronunciation katakana notation knowledge that stores knowledge for writing the pronunciation obtained by the pronunciation Japanese translation unit 5 in katakana. Base 6 and this pronunciation katakana notation knowledge base 6
A pronunciation katakana notation unit 7 that searches for the word and derives the notation when the word is written in katakana, and a plurality of possible Japanese words corresponding to the input foreign language word based on the output of this pronunciation katakana notation unit 7. The Japanese notation display section 8 outputs the notation.

Next, the operation of this embodiment will be explained.

[0019] Basically, this embodiment can be applied to spelling conversion from any foreign language as long as it is spelled with phonetic characters, but here we will convert from English word spelling to Japanese katakana spelling. This will be explained using conversion as an example.

The word list input unit 1 is a character input means such as a keyboard. The spelling of the word in the original language input from here is sent to the foreign language pronunciation analysis section 3. The foreign language pronunciation analysis unit 3 searches the foreign language pronunciation analysis knowledge base 2 and breaks down the input word into units for pronunciation estimation. Then, these decomposed units are combined to estimate the pronunciation of the entire input word.

As for the analysis method of the foreign language pronunciation analysis knowledge unit 3 and the foreign language pronunciation analysis knowledge base 2, already known methods for text-speech conversion can be used. For example, Honey S. Elovitz (Hone
yS. Elovitz et al.
IEEE Transactions on Acoustics Speech and Signal Processing
on Acoustics, Speech, and
The paper “Letter to Sound Rules” published in the journal SignalProcessing
Four automatic translations
of English Text to Phonetics (Letter-to-Sound
Rules for Automatic Tr
onslation of English T
The method of converting the spelling of a word into its pronunciation, such as the one described in "Ext to Phonetics", can be directly adopted.

[0022] The pronunciation Japaneseization section 5 includes the foreign language pronunciation analysis section 3.
After receiving the spelling of the original word and the estimated pronunciation from
Using the knowledge stored in the Japanese pronunciation knowledge base 4, the pronunciation of the input word in its original language is made to correspond to the Japanese pronunciation.

As the knowledge to be stored in the pronunciation Japaneseization knowledge base 4, already known knowledge regarding the Japaneseization of foreign languages can be used. For example, the rules of change when a foreign language becomes Japanese, as described in the book ``Foreign Language Formation and Its Education'' published by the National Institute for Japanese Language and Linguistics by Hiroko Cuckenbusch and Mieko Oso, can be used as is. Can be adopted.

When there are multiple pronunciation possibilities of the original language translated into Japanese, the pronunciation Japanese pronunciation section 5 outputs all of them.

The Japaneseized pronunciation is transmitted to the pronunciation katakana notation unit 7, where it is transferred to the pronunciation katakana notation knowledge base 6.
is converted to katakana based on the knowledge stored in . At that time, if there are multiple possibilities for the katakana notation representing the pronunciation, all of them are output.

The Japanese notation display unit 8 outputs and displays the katakana notation of the input word obtained by the pronunciation katakana notation unit 7.

FIG. 2 is a flowchart showing the operation of the foreign language pronunciation analysis section shown in FIG. 1, FIG. 3 is a flowchart showing the operation of the pronunciation Japaneseization section shown in FIG. It is a flowchart which shows the operation|movement of a notation part.

Next, the operation of this embodiment will be explained based on a specific example with reference to FIGS. 1 to 4.

In FIG. 2, the foreign language pronunciation analysis knowledge base 2 holds contents in the format shown in Table 1.

[0030]

For example, when the English word "face" is input through the word table input section 1, the foreign language pronunciation analysis section 3 searches the foreign language pronunciation analysis knowledge base 2 and finds the two pronunciation units /fa/ce/. By decomposing the words and combining their pronunciations, it is inferred that the pronunciation of the entire word is [feis].

In FIG. 3, pronunciation Japanese knowledge base 4
Contents in the format shown in Table 2 are held within.

[0033]

For example, upon receiving [feis] as the output result of the foreign language pronunciation analysis unit 3, the pronunciation Japanese pronunciation unit 5 searches the pronunciation Japanese pronunciation knowledge base 4 and converts the original language pronunciation into Japanese. let In this example, the vowel [ei] is sometimes translated into Japanese as [ei] and sometimes becomes [e:], so the resulting Japanese pronunciation of the word is [
There are two types: feisu] and [fe:su].

FIG. 4 is a flowchart showing the operation of the pronunciation katakana notation unit 7.

The pronunciation katakana notation knowledge base 6 holds contents in the format shown in Table 3.

[0037]

For example, the output result of the pronunciation Japanese conversion unit 5 [
feisu], [fe:su], the pronunciation katakana notation 7 searches the pronunciation katakana notation knowledge base 6, divides it into units of Japanese notation corresponding to the Japanese pronunciation, and returns [f
/fe/ for [eisu] and [fe:su] respectively.
The parts i/su/, /fe:/su/ are obtained. These are made to correspond to katakana, but in this example, there are two possibilities for the katakana notation of the beginnings of the word /fe/ and /fe:/: fe and hue, and fe and hue. Therefore, it can be seen that there are four possibilities for the katakana notation of the entire input word: face, face, face, and face.

As described above, the spelling of a foreign language word written in phonetic characters can be converted into katakana. Even when multiple katakana notations are possible, they can be output. The results thus obtained are sent to the Japanese notation display section 8 and displayed there.

[0040]

[Effects of the Invention] As explained above, the present invention provides a word list input means for inputting foreign words written in phonetic characters;
A foreign language pronunciation analysis knowledge base that stores knowledge for deriving the pronunciation of the input word from the notation of the word input using this word table input means, and the input word by searching this foreign language pronunciation analysis knowledge base. A foreign language pronunciation analysis means that guides the pronunciation of the word, a pronunciation Japaneseization knowledge base that stores knowledge for estimating the pronunciation when the foreign word is translated into Japanese from the pronunciation and transcription of the foreign word, and this A pronunciation method for Japanese pronunciation that searches the pronunciation knowledge base and guides the pronunciation when the input word is translated into Japanese, and knowledge for writing the pronunciation obtained by this pronunciation Japanese method in katakana. a stored pronunciation katakana notation knowledge base; a pronunciation katakana notation means that searches this pronunciation katakana notation knowledge base and guides the notation when the input word is written in katakana;
By having a Japanese notation display means that outputs a plurality of possible Japanese notations corresponding to the input foreign language word based on the output of this pronunciation katakana notation means, when creating a dictionary or database heading. Since the possibility of variations in spelling when writing a foreign language in Japanese katakana is automatically output, the number of man-hours required to come up with possible spelling variations is greatly reduced, resulting in better headlines. This has the effect of greatly reducing search omissions caused by different spellings when searching the dictionary or database.

[Brief explanation of drawings]

FIG. 1 is a block diagram showing a foreign language katakana notation conversion support device according to an embodiment of the present invention.

FIG. 2 is a flowchart showing the operation of the foreign language pronunciation analysis section shown in FIG. 1;

FIG. 3 is a flowchart showing the operation of the Japanese pronunciation unit shown in FIG. 1;

FIG. 4 is a flowchart showing the operation of the pronunciation katakana notation unit shown in FIG. 1;

[Explanation of symbols]

1 Word table input section 2 Foreign language pronunciation analysis knowledge base 3 Foreign language pronunciation analysis section 4 Japanese pronunciation knowledge base 5 Pronunciation Japanese translation section 6 Pronunciation Katakana notation knowledge base 7 Pronunciation Katakana notation section 8 Japanese notation display section

Claims (1)

[Claims] 1. A word table input means for inputting foreign words written in phonetic characters, and knowledge for deriving the pronunciation of the input word from the notation of the word input by the word table input means. A stored foreign language pronunciation analysis knowledge base, a foreign language pronunciation analysis means that searches this foreign language pronunciation analysis knowledge base to derive the pronunciation of the input word, and a foreign language pronunciation analysis means that searches the foreign language pronunciation analysis knowledge base to derive the pronunciation of the input word, A pronunciation Japanese knowledge base that stores knowledge for estimating the pronunciation when the word is translated into Japanese, and this pronunciation Japanese translation knowledge base is searched to find the pronunciation when the input word is translated into Japanese. A pronunciation Japaneseization means that guides the pronunciation, a pronunciation katakana notation knowledge base that stores knowledge for displaying the pronunciation obtained by this pronunciation Japaneseization means in katakana, and a pronunciation katakana notation knowledge base that searches for the inputted pronunciation. A pronunciation katakana notation means that guides the notation when a word is written in katakana, and a Japanese language device that outputs a plurality of possible Japanese notations corresponding to the input foreign language word based on the output of the pronunciation katakana notation means. 1. A foreign language katakana notation conversion method characterized by having a notation display means.