JPH05342251A - Dictionary retrieval device - Google Patents

Abstract

PURPOSE:To easily obtain a character string including KANJI (Chinese character) being a correct output by retrieving a word dictionary with an auxiliary character string and an input character string. CONSTITUTION:A code conversion part 2 is provided with a conversion table storage part 4 storing a character which is difficult to discriminate in pronunciation in accordance with a character different in the same pronunciation or similar pronunciation. When the character which is difficult to discriminate in pronunciation is included in an input KANA (Japanese syllabary) character string stored in the temporary storage device 1, the character is substituted for the other character and code conversion is executed. In such a case, the conversion table storage part 4 stores the character codes which are difficult to clearly discriminate on pronunciation by making them to mutually correspond each other. When the character which is difficult to discriminate on pronunciation is included in the input KANA character string, a dictionary consultation control part 3 retrieves the word dictionary as against the input KANA character and retrieves the word dictionary as to the auxiliary KANA character.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a dictionary search device in a kana-kanji conversion device, and particularly to a dictionary search device capable of performing an appropriate conversion process on an input character string containing a kana character which is difficult to distinguish by pronunciation. Regarding

[0002]

2. Description of the Related Art Recently, sentence creating devices such as Japanese word processors have become popular. This kind of device generally inputs a kana character into a created sentence and converts it into a character sequence containing mixed kanji for each word to create a Japanese sentence. This kana / kanji conversion is performed by searching a character string, which is registered in advance in the word dictionary and includes kanji corresponding to each kana character string. That is, a character string including Kanji is provided corresponding to the kana character string. Therefore, if the kana character string is erroneously input, naturally, the kanji conversion is not performed, or the wrong kanji conversion character string is output.

[0003]

However, many Japanese words include characters that cannot be distinguished phonetically, such as "zu" and "zu", "ji" and "ぢ". In this case, if the kana character string is not entered correctly, incorrect Kanji conversion will be performed. For example, in the case of not creating a sentence "build", it is necessary to input "Kizuku" as a kana character string, but at this time, when "Kizuku" is input, a Kanji conversion output of "Aware" is obtained. Therefore, in such a case, it is necessary to correct the sentence by re-inputting the kana character string. Further, in recent years, text editing by voice input has been directed. This device analyzes phonemes input by voice and converts them into phonetic character strings, and then converts them into kanji characters to form sentences.At this time, however, there is a problem in handling the above-mentioned characters that cannot be distinguished in terms of pronunciation. Become. Similarly, many Japanese words include a character whose pronunciation is similar because a kana containing a vowel "o" in pronunciation is followed by a kana "o" or "u". For example, if you want to get the word "prince", you need to enter "Ouji" as a kana character string, but if you mistakenly enter "Oji" at this time, you will get a Kanji conversion output such as "Oji". , You must re-enter the kana character string. Even when inputting this by voice, "Ouji" and "Oji" can be distinguished by pronouncing them one by one, but when they are pronounced consecutively, the sounds are similar, and such pronunciation is similar. The same problem as above occurs for characters.

The present invention has been made in consideration of such circumstances, and an object thereof is to obtain a character string containing a kana character which is difficult to distinguish in pronunciation, It is to provide a highly practical dictionary search device capable of easily finding an appropriate kanji conversion output.

[0005]

[Means for Solving the Problems] In the present invention, the pronunciation is "O".
When an input string containing the kana characters "o" or "u", which are difficult to distinguish because the kana containing the vowels are followed by the kana "o" or "u", this kana is The letters "o" and "u" are replaced with "u",
By generating an auxiliary character string replaced with "O" and searching the word dictionary with both this auxiliary character string and the input character string, it is possible to easily obtain a character string containing kanji that is the correct conversion output. It was done.

[0006]

According to the present invention, even a word including "O" or "U" following a vowel "O" which is difficult to distinguish in pronunciation, either "O" or "U" (conventional It is possible to perform appropriate Kanji conversion processing simply by inputting characters that are not considered to be correct.

[0007]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of the present invention will be described below with reference to the drawings.

FIG. 1 is a schematic configuration diagram showing a main part of the apparatus of the embodiment. The kana character string used for writing a sentence is character-coded through an input device (not shown), and is subjected to preprocessing such as division into words, and then input. The input device may be a keyboard device such as a kana keyboard or a voice input recognition device. The input kana character string thus input is temporarily stored in the temporary storage device 1 and then given to the dictionary lookup control unit 3 via the code conversion unit 2. The code conversion unit 2 includes a conversion table storage unit 4 that stores characters that are difficult to distinguish by pronunciation in correspondence with different characters that have the same or similar pronunciation, and is stored in the storage device 1. When the input kana character string contains a character that is difficult to distinguish by the above pronunciation, the character is replaced with another character for code conversion. The auxiliary kana character string generated by this code conversion is also given to the dictionary lookup control unit 3 in the same manner as the input kana character string.

In other words, the conversion table storage unit 4 produces pronunciations such as "ji" and "ji", "zu" and "zu", and "o" and "u".
Character codes that are difficult to clearly distinguish between the two are stored in correspondence with each other. In this way, for example, "Kizuku"
When an input kana character string of is given, an auxiliary kana character string of "kizuku" is generated. Similarly, when an input kana character string "Oji" is given, an auxiliary kana character string "Ouji" is generated.

On the other hand, the dictionary storage unit 5 which is a word dictionary stores and registers a plurality of words by associating a word consisting of a character string including Chinese characters with a kana character string indicating its pronunciation. These word information is registered and set in advance,
For example, the addresses are sequentially assigned and registered from the item "a" to the item "n". With respect to such a dictionary storage unit 5, the dictionary lookup control unit 3 first sequentially collates the given input kana character string with the kana character string in the dictionary, and detects the matching. Then, the character string having the kanji error in the word dictionary in which the match is detected is selectively read. At this time, if the input kana character string does not include characters that are difficult to distinguish by pronunciation, the selected kanji mixed character string is output as it is as kanji conversion output data. .. If the input character string does not include a character that is difficult to distinguish by pronunciation and the corresponding word cannot be found in the word dictionary, the input kana character string is output as it is. On the other hand, when the input kana character string contains a character that is difficult to distinguish by pronunciation, the dictionary lookup control unit 3 first searches the word dictionary for the input kana character string, and then creates the character code by conversion. The word dictionary is searched for the auxiliary kana character string. Then, the conversion output composed of the mixed Kanji characters obtained by these both searches is obtained as a conversion candidate. Then, the correct one of the conversion candidate outputs is selected by, for example, displaying it on the display by an operator's instruction, or by performing a context analysis, or by referring to information on the appearance frequency in the sentence. Done.

According to the present apparatus thus constructed, the auxiliary kana character string "Kizuku" is generated for the input kana character string in which "Kizuku" is input, and these character strings and the word dictionary are generated. The conversion output of “build” and “notice” is obtained by collating with, and an appropriate one of them is selected and output. In addition, a conversion output of "nosebleed" can be obtained by the auxiliary kana character string of "hanaji" with respect to the input kana character string erroneously input as "hanaji". Similarly, for the input kana character string "majika", the kanji conversion output "close" can be obtained by the auxiliary kana character string "majika". Furthermore, an auxiliary kana character string "Ouji" is generated for the input kana character string in which "Oji" is input, and by matching these character strings with the word dictionary, conversion candidates such as "Oji" and "Oji" are generated. Then, an appropriate one selected by the context analysis is selected as the conversion output from among them.

In this way, in this apparatus, even if a character string including a character that is difficult to distinguish in terms of pronunciation is given, this can be effectively converted into a correct Kanji character. Therefore, mistakes in writing sentences due to operator's mistakes in key entry or not knowing customary spelling,
The troublesomeness of performing the calibration work for this is eliminated. Moreover, it is possible to sufficiently deal with the text creation by voice input, and the effect is very large.

The present invention is not limited to the above embodiment. For example, character strings with similar pronunciations (for example, "Oji" and "Ouji") may be registered in the word dictionary in association with each other, and the input character strings may be searched for at the same time. .. After the dictionary is collated once with the input character string, if the corresponding word does not exist or the conversion output is improper, the control sequence is changed so that the character code is converted and the dictionary is collated again. May be set. In short, the present invention can be modified and carried out without departing from the scope of the invention.

[0014]

As described above in detail, according to the present invention, even if an operator inputs an incorrect kana character or a kana character string given by voice input, the kanji character is always converted accurately. This has a great practical effect such as the ability to create Japanese sentences.

[Brief description of drawings]

FIG. 1 is a schematic configuration diagram showing a main part of an apparatus according to an embodiment of the present invention.

[Explanation of symbols]

1 ... Temporary storage device, 2 ... Code conversion unit, 3 ... Dictionary lookup control unit, 4 ... Conversion table storage unit, 5 ... Dictionary storage unit.

 ─────────────────────────────────────────────────── ─── Continuation of the front page (72) Hiromi Saito, 1 Komukai Toshiba-cho, Sachi-ku, Kawasaki-shi, Kanagawa, Ltd. Toshiba Research Institute Co., Ltd. Bashi Stock Company, Toshiba Research Institute

Claims (1)

[Claims] 1. A word dictionary in which a plurality of words including Chinese characters are stored together with a character string representing the pronunciation, a storage device for temporarily storing an input kana character string, and the storage device stored in the storage device. In the input kana character string,
When the pronunciation includes the kana character "o" or "u", which is similar in pronunciation because the kana character including the vowel "o" is followed by the kana character "o" or "u", the kana character "o", Means for generating an auxiliary kana character string in which “u” is replaced with “u” and “o”, and dictionary matching means for searching the word dictionary with the input kana character string and the auxiliary kana character string, respectively. A dictionary search device comprising: