JPS6126172A - Kana/kanji conversion system - Google Patents

Abstract

PURPOSE:To execute homophone selecting processing using a term dictionary efficiently and to reduce the volume of the word dictionary by registering terms to be described in each term information in a word dictionary and adopting only words having at least one homophone. CONSTITUTION:A ''kana'' (Japanese syllabary) character string inputted from a ''kana'' character string input means 2 is converted into a word network by an analysis means 4 on the basis of the retrieval of the word dictionary 3 and the inspection of grammatical connection and the word network is sent and stored to/in a network storing means 5. A ''kanji'' (Chinese character/''kana'' mixed character string output means 7 forms a ''kanji''/''kana'' mixed character string on the basis of the formed word network while referring to the term dictionary 6 and outputs the formed character string to a display output means 8. The term dictionary 6 is referred only when branching is generated because of the presence of homophones on the word network. Namely, words to be described in each word information are registered in the word directionary 6 and only words having at least one homophone are adopted. Thus, the homophone selecting processing using the term dictionary can be efficiently executed and the volume of the term dictionary 6 can be reduced.

Description

DETAILED DESCRIPTION OF THE INVENTION (Field of Industrial Application) The present invention relates to a kana-kanji conversion method that inputs a kana character string and outputs a character string containing kanji and kana.

(Prior art and its problems) In recent years, the kana-kanji conversion method, which is one of the Japanese input methods, has been widely adopted in Japanese input information processing devices such as Japanese word processors.

In kana-kanji conversion, so-called homophone processing, which obtains the desired result from a string of multiple kanji and kana characters corresponding to one kana string, is one of the very important elements, and various methods have been used to date. Several proposals have been made, and some of them have been put into practical use. One of such homophone processing methods is the use of terminology information. This is a method in which compound words and idiomatic expressions are registered in a term dictionary in advance, and when selecting homophones, the priority of the compound words and idiomatic expressions registered in the term dictionary is increased. For example, if you store terminology information in the terminology dictionary that shows that there is a relationship between ``kanji'' and ``conversion'' and ``territory'' and ``return'' as shown in Figure 1, then you can Even if the homonyms ``conversion'' and ``return'' occur, ``kanji conversion'' can be obtained from ``kanjihenkan,'' and ``return the territory'' can be obtained from ``shiyodo wo henkansuru.'' . In addition, as a document regarding the use of terminology information, for example, Information Processing Society of Japan Journal Vol.

Examples include ``2241r Kana-Kanji Conversion System for Solid Text and Its Homophone Processing'' and JP-A-59-2125 ``Kana-Kanji Conversion System''.

In this way, a term dictionary that is effective in the homophone selection process in kana-kanji conversion is usually created by extracting frequently used examples from general documents and then tabulating them. At this time, it is created regardless of the word weight in the word dictionary used for kana character string analysis in kana-kanji conversion.
Even if the term is very commonly used,
If the word is related to a word that is not registered in the word dictionary, what is used in the homophone selection process is, for example, if the information that "tamaishi" and "mixed circulation" are related is registered in the word dictionary. However, if the kana-kanji conversion from the kana character string "gyokuseki konkou" does not result in "tamaishi" and "mixed water", this terminology information will not be referenced. Also,
Even if the term is very frequently used, this term information is not used even if no homophones are generated as a result of kana-kanji conversion. Examples of this include "Instruction" → "Research" and "Development".

In this way, the terminology dictionary that has been used in the past, which simply stores terminology information such as compound words and idiomatic expressions in Japanese, cannot be used for homophone selection processing in kana-kanji conversion). This resulted in a large increase in memory capacity.

(Object of the Invention) An object of the present invention is to eliminate the drawbacks of the prior art as described above and to provide a kana-kanji conversion method having a term dictionary suitable for homophone selection processing for kana-kanji conversion. .

(Structure of the Invention) According to the present invention, there is provided an analysis means for creating a word network by repeatedly performing word dictionary searches and grammatical connection tests based on input kana character strings; A term dictionary that stores terminology information describing combinations of words used in compound words, idiomatic expressions, etc., and a kanji character string that forms and outputs a probable kanji-kana-mixed character string from the word network by referring to the term dictionary. In a kana-kanji conversion method that includes at least a means for outputting kana-mixed p character strings, the term information is configured only for words existing in the word dictionary, and each of the term information includes homophones among the words. It is possible to obtain a kana-kanji conversion method characterized by including at least one of the following.

(Detailed Description of Configuration) The present invention eliminates the drawbacks of the prior art by adopting the above configuration. First, by configuring the terminology information to be stored in the terminology dictionary only for words that exist in the word dictionary used in kana-kanji conversion, it is possible to It is possible to remove from the term dictionary information on terms that are never referenced even if the term is used. Furthermore, at least one word that has a homophone is included in the word group whose relationship is described in each terminology information (in other words, all the word groups whose relationship is described in a certain terminology information do not have a homophone). By not placing this term information in the term dictionary, you can remove from the term dictionary the term information that is wasted because no homophones occur during the word dictionary search and connection test in kana-kanji conversion. As a result, only terminology information that could be used in the homophone selection process was included. A term dictionary with high storage efficiency can be provided.

(Embodiments) Hereinafter, embodiments of the present invention will be described in detail with reference to all the drawings.

FIG. 2 is an embodiment of the present invention. 1 is a control means for overall control, 2 is a keyboard, etc.

This is a kana character string input means that takes the kana character string and sends it to the analysis means described later. 3 is a word dictionary that contains the pronunciation of words, kanji notation, 9 parts of speech information, etc. required for kana-kanji conversion. 4 is an analysis means for creating a word network to be described later while repeating searches in the word dictionary 3 and grammatical connection tests based on the kana character string sent from the kana character string input means 2. 5 is a network storage means for storing the word network created by the analysis means 4, and 6 is a network storage means for storing word networks used in Japanese compound words and idiomatic expressions. 7 is a term dictionary that stores terminology information, and 7 is a terminology dictionary that stores terminology information, and 7 is a terminology dictionary that traces likely branches on the word network in the network storage means 5 by referring to the terminology information in the terminology dictionary 6. , is a kanji/kana mixed 9 character string output means for creating and outputting a kanji/kana mixed character string, and 8 is a display output means, such as a CRT, for displaying and outputting the obtained kanji/kana mixed character string. Further, FIG. 3 is a diagram conceptually showing the storage status of the word dictionary 3. Furthermore, Fig. 4 is a diagram conceptually showing the storage status of the terminology dictionary that has been used in the past), and Fig. 5 shows examples of terminology used to create the terminology dictionary shown in Fig. 4. 4 is a diagram conceptually showing the storage status of the term dictionary 6 when the example of the word dictionary 3 shown in FIG. 3 is used. As can be seen by comparing Figures 4 and 5, "Northern Territory" and "Ryodo - Expedition" which exist in Figure 4 do not exist in Figure 5. This is because "northern" and "territory" are both terms that do not have homophones in Figure 3, and "Ryodo" does not exist in Figure 3, so it is not adopted as the terminology information shown in Figure 5. It is from.

Here, the processing procedure in this embodiment when the kana character string "Happoshiyoudohenkan" is input will be described. first,
The kana character string inputted by the kana character string input means 2 is converted into a word network by searching the word dictionary 3 and testing grammatical connections in the analysis means 4, and is sent to the network storage means 5. The word network is a network structure that views five Japanese sentences as a series of words that can be connected grammatically. Figure 6 shows an example of a word network created from "Happoshi Yodo Henkan". When using the example in Figure 3,

The node for "Ryodo" indicated by a broken line on the network is not created because "Ryodo" is not registered in the dictionary.
As a result, a network connected by solid lines is stored in the network storage means 5.

The kanji/kana mixed character string output means 7 creates a kanji/kana mixed character string based on the created word network while referring to the term dictionary 6, and outputs it to the display output means 8.

In the example in Figure 5, the relationship "territory-return" is described, so "return" is adopted at the branch of "conversion" and "return" in the word network shown in Figure 6, and the final word is "return."'
The return of the Northern Territories will be obtained.

Examples of the processing procedures of the analysis means 4 and the kanji-kana mixed character string output means 7 are shown in FIGS. 7 and 8, respectively, and these are internal processing procedures that have been put to practical use in Japanese word processors that input kana-kanji conversion input. are essentially the same thing.

As shown in the example above, the term dictionary 6 is referred to only when a branch occurs due to the presence of a homophone on the word network. the target word.

``All of them are registered in the word dictionary.

and at least one of which has a homophone.

(Effects of the Invention) As described above, by using the present invention, the capacity of the term dictionary in the kana-kanji conversion method can be reduced.

[Brief explanation of the drawing]

Figure 1 is a configuration diagram showing an example of term information in a term dictionary;
Fig. 2 is a block diagram for explaining one embodiment of the present invention, Fig. 3 is a diagram conceptually showing the storage status of a word dictionary, and Fig. 4 is a storage diagram of a term dictionary conventionally used. FIG. 5 is a diagram conceptually showing the storage situation of the term dictionary created in the present invention. FIG. 6 is a configuration diagram showing an example of a word network. FIG. 7 8 is a flowchart showing an example of the processing procedure of the analysis means 4, and FIG. 8 is a flowchart showing an example of the processing procedure of the character string outputting means 7 containing Kanji and kana. In the figure, 1... control means, 2... kana character string input means. 3... Word dictionary, 4... Analysis means. 5. Network storage means. 6...Term dictionary. 7. A means of outputting character strings containing kanji and kana. 8.Display output means. Figure 1 Figure 3 Figure '''7 Figure 6

Claims (1)

[Claims] An analysis method that creates a word network by repeatedly searching word dictionaries and grammatical connection tests based on the input kana character string, and an analysis method for words used in Japanese compound words and idiomatic expressions. Kana-Kanji conversion comprising at least a term dictionary that stores terminology information describing combinations, etc., and a means for outputting a string of characters mixed with Kanji and Kana, which forms and outputs a string of characters mixed with Kanji and Kana from the word network by referring to the term dictionary. In the method, the terminology information is composed only of words existing in the word dictionary, and each terminology information includes at least one of the words that has a homophone. Kana-kanji conversion method.