JPH08212210A - Document generation device, kana(japanese syllabary) mixed with kanji(chinese character)/kanji conversion method and dictionary generation device and its method - Google Patents

Abstract

PURPOSE: To generate and output conversion candidates obtained by making a user to convert a part which is KATAKANA(square form of KANA)-inputted intending KANJI notation into KANJI and convert a part which is KANA-inputted intending KANA notation into KANA without increasing the capacity of a dictionary, and to improve the conversion efficiency of a KANA mixed with KANJI character string. CONSTITUTION: The dictionary 113 has information on the correspondence of KANJI of headwords and reading. Only words having the most KANJI notation parts in the same word different in notation are registered. A conversion part 108 retrieves the dictionary 113, obtains conversion candidate information on the character string where KATAKANA is converted into KANA and judges the connection of conversion candidate information. A character string generation part 111 substitutes the KATAKANA part of the inputted character string for KANJI or KATAKANA or the HIRAGANA part of the inputted character string for HIRAGANA from the headwords registered in the dictionary 113 of the conversion candidate whose connection judgement is permitted so as to output it.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a document creation device for creating a document containing kanji characters by inputting a kana-kanji character string by, for example, a handwriting input method, and more particularly to a document creation device characterized by conversion candidate creation processing. The present invention relates to a device and a kanji conversion method for mixing kanji, and a dictionary creating device and a dictionary creating method used in this document creating device.

[0002]

2. Description of the Related Art Conventionally, there is a document creating apparatus such as a Japanese word processor that creates a kanji mixed document by inputting a kana-mixed kana character string by, for example, a handwriting input method. In this type of document creation device, the dictionary used for kana-kanji conversion stores information about the correspondence between each kanji in the heading and the reading.
The same word but different notation, such as "Suzumushi", were registered separately in the dictionary.

Further, in the conversion, only the notation registered in the conversion dictionary can be converted. Therefore, in order to obtain conversion candidates for several notations for one word, it is necessary to register the dictionary for each notation.

[0004]

As described above, in the conventional document creating apparatus, in order to output the conversion candidates of the notation intended by various users such as the Kanji notation or the Hiragana notation, the respective candidates are output. It was necessary to register in the dictionary for each notation. For this reason, there is a drawback that the capacity of the dictionary becomes huge.

The present invention has been made in consideration of the above circumstances, and intends to use Hiragana notation in which the user inputs katakana in Kanji without increasing the capacity of the dictionary. A document creation device and a kanji-mixed kanji conversion method that can improve the conversion efficiency of kanji-mixed kana character strings by creating and outputting a conversion candidate in which the input hiragana is written in hiragana, and a dictionary used in this document creation device. An object is to provide a creating device and a dictionary creating method.

[0006]

SUMMARY OF THE INVENTION According to the present invention, in a document creating apparatus having a kanji / kanji / kanji conversion function, input knot / kanji storage means for storing inputted kanji / kanji / kana character string information, and this input When the input character string information stored in the notation storage means includes katakana, it has notation conversion means for converting the katakana to hiragana and information on the correspondence between each kanji and heading of the heading, and the same notation Kanji-kana-kanji conversion dictionary in which only the words with the most kanji-writing parts are registered, and this kanji-kana kana-kanji conversion dictionary are searched for the character string information in which katakana is converted to hiragana by the above notation conversion means. A kanji-kana kana-kanji conversion device that obtains conversion candidate information and determines whether the conversion candidate information is connected, and a kanji-kana kana-kanji converter. For conversion candidate information for which connection determination is permitted by the above, replace the katakana part of the input character string information with kanji or katakana or the hiragana part of the input character string information from the heading registered in the Kanji-mixed kanji conversion dictionary. According to the configuration, a candidate character string generation unit including kanji and katakana that creates the candidate character string information of the notation according to the conversion instruction in the input is provided (first configuration).

[0007] In addition, there is information about the correspondence between each kanji of the heading and the reading, and only the word with the most kanji notation part is registered in the same word with different notation, and the hiragana notation of the registered heading is registered. Search the kanji-mixed kana-kanji conversion dictionary in which markers are added to possible kanji and this kanji-mixed kana-kanji conversion dictionary, obtain conversion candidate information for the input kanji-mixed kana character string information, and determine the connection of the conversion candidate information For Kanji-mixed kana-kanji conversion means for performing and conversion candidate information for which connection determination is permitted by this kanji-mixed kana-kanji conversion means, the part of the input character string information that can be written in kanji is registered from the headline registered in the kanji-mixed kana-kanji conversion dictionary. All in Kanji, and
According to a configuration including a kanji-blended candidate character string generation means for creating candidate character string information in which the hiragana notation portion provided with the marker information is written in hiragana (second configuration).

Further, in the first and second configurations, depending on whether or not the input character string information includes katakana, whether the kanji / katakana mixed candidate character string means or the kanji mixed candidate character string generation means is used. A conversion determination means for determining is provided (third configuration).

The dictionary creating device of the present invention detects a same word with different kanji in a kanji-kana-kanji conversion dictionary and a kanji-kanji kana-kanji conversion dictionary having information relating to the correspondence between each kanji of a headline and reading. Means and marker information is added to the kanji written in hiragana in other notation words for the most common kanji notation part among the same notations detected by this same word detection means, and other than that. It is characterized by comprising a same word merging means for deleting, and creating a kanji-kanji conversion dictionary in which only kanji characters having the largest number of kanji characters are registered among the same words having different notations.

[0010]

According to the first configuration, a mixture of Kanji characters, which has information about the correspondence between each Kanji character in the heading and the reading, and in which only the words with the most Kanji notation part among the same words with different notations are registered. A kana-kanji conversion dictionary is used to replace the katakana part of the input character string information with kanji or katakana, or replace the hiragana part of the input character string information with hiragana to create candidate character string information that corresponds to the conversion instruction during input. To be done.

According to the above-mentioned second structure, since there is information on the correspondence between each kanji of the heading and the pronunciation, only the word having the most kanji notation part is registered among the same words having different notations. A kana-kanji conversion dictionary is used in which marker information is added to the kanji characters that can be written in hiragana in the specified heading, and all kanji-expressible parts of the input character string information are converted to kanji. The candidate character string information in which the kanji part that can be written in hiragana with the marker information is written in hiragana is created.

According to the third configuration, depending on whether or not the input character string information includes katakana, when the input character string information includes katakana, the headline registered in the kanji-kana-kanji conversion dictionary. By replacing the katakana part of the input character string information with kanji or katakana or replacing the hiragana part of the input character string information with hiragana, the candidate character string information of the notation according to the conversion instruction at the input is created as a conversion candidate. If the input character string information does not include katakana, all the characters that can be written in kanji in the input character string information from the heading registered in the kanji / kanji conversion dictionary are written in kanji, and marker information is added. The candidate character string information in which the kanji part that can be written in hiragana is written in hiragana is created.

Further, in the dictionary creating apparatus of the present invention,
If there are multiple registered words for each heading that has a different notation in the kanji-mixed kana-kanji conversion dictionary that has information about the correspondence between each kanji in the heading and the reading, the hiragana is written in one of the registered headings. Marker information is added to existing Kanji characters, and only words with the most Kanji characters are left and other words with different Kanji characters are deleted. A kanji conversion dictionary with mixed kanji is created.

[0014]

An embodiment of the present invention will be described below with reference to the drawings. FIG. 1 is a block diagram showing an embodiment of a document creating apparatus of the present invention. The input device 101 is a device that can input a kana-kanji character string by recognizing a handwritten character from coordinates obtained from a handwriting tablet or the like and outputs input information as a character code.

The display device 119 is a CRT (Cathode Ray).
It is a screen display device such as a Tube) or LCD (Liquid Crystal Display), and a user can interactively create a document by using these.

The input control unit 102 temporarily buffers the character data input from the input device 101 in the input buffer 103 while waiting for conversion processing, creates a character string mixed with Kanji, and sends it to the notation conversion unit 105.

The notation conversion unit 105 includes an input notation storage unit 10
6, input character string is entered and input notation saving buffer 107
In addition, when the input character string includes katakana, the katakana in the input character string is replaced with hiragana and sent to the kanji / kanji / kanji conversion unit 108.

The kanji-mixed kana-kanji conversion dictionary 113 has information about the correspondence with each kanji of the headline, and only the words having the most kanji notation parts among the same words with different notations are registered, and kanji that can be written in hiragana. A marker is attached to.

The kanji-mixed kana-kanji conversion dictionary 108 temporarily stores the character string in the conversion buffer 109 while the character string is an appropriate conversion target such as a word or a phrase, and after the character string is converted, the kanji-mixed kana-kanji conversion dictionary 113. To search. Next, the connection determination of each conversion candidate that has been successfully searched is performed, and the connectable conversion candidates are sent to the conversion determination unit 110.

The conversion determination unit 110 refers to the input notation storage buffer 107 to check whether or not the input character string corresponding to the conversion candidate includes katakana. If the input character string includes katakana, the candidate characters mixed with kanji and katakana are included. Column generator 1
If 11 does not include katakana, the conversion candidates are sent to the kanji mixed candidate character string generation unit 112.

Candidate character string generation unit 11 containing kanji and katakana
1 refers to the kana-mixed kana-kanji conversion dictionary 113, and checks whether the part input in katakana can be written in kanji or katakana, or the part input in hiragana can be written in hiragana. If possible, katakana was input. Create a candidate character string so that the part is kanji or katakana and the input part is hiragana.

The kanji-mixed candidate character string generation unit 112 refers to the kanji-mixed kana-kanji conversion dictionary 113 to indicate a candidate character string in which all kanji-expressible parts are written in kanji, and a marker indicating that hiragana can be written. If there is a given part, create a candidate character string with that part in hiragana notation.

When the kanji-mixed kana-kanji conversion unit 108 finishes creating the candidate character string, it sends the conversion candidates to the conversion candidate control unit 114 and stores them in the candidate buffer 115. The conversion candidate control unit 114 determines the priority according to the number of connection points, the frequency, the learning state, etc., and rearranges the conversion candidates stored in the candidate buffer 115 in order of priority. Then, the candidate character string is displayed to prompt the user to select a candidate, and the candidate character string selected by the user is stored in the document buffer 116.

The display control unit 117 includes a document buffer 116.
Develops the data stored in the display buffer 118 in accordance with the format set in advance by the format setting unit 104,
It is sent to the display device 119.

In the first configuration of the present invention, the conversion determination unit 110 and the Chinese character mixture candidate character string generation unit 112 are unnecessary, and in the second configuration of the present invention, the notation conversion unit 105 and the input notation storage unit. 106, the input notation storage buffer 107, the conversion determination unit 110, and the Kanji / Katakana mixed candidate character string generation unit 111 are unnecessary.

Next, the operation of the embodiment will be described. 2 and 3 are flowcharts showing the flow of processing from character input to selection of conversion candidates in the embodiment of the document creation apparatus of the present invention. The character data input from the input device 101 is sent from the input control unit 102 to the notation conversion unit 105. This input data is stored in the input notation storage buffer 107 by the input notation storage unit 106 in the state of the input notation (steps 210 and 201).

Further, the notation conversion unit 105 replaces katakana with hiragana of the input character and stores it in the conversion buffer 109 (steps 203 to 205).

The kana / kanji / kanji conversion unit 108 monitors the conversion buffer 109, and repeats the processes from the character input to the storage of the conversion buffer 109 until it becomes a conversion target (steps 201 to 206).

When the character string stored in the conversion buffer 109 is to be converted, the kanji / kanji / kanji conversion unit 108 retrieves conversion candidates from the kanji / kanji / kanji conversion dictionary 113 (step 207). When the dictionary search is completed, connection determination of all conversion candidates that have been successfully searched is performed (step 208).

Then, with respect to the conversion candidates for which the connection judgment is permitted, the conversion judgment unit 110 causes the input notation storage buffer
7, it is checked whether the input character string corresponding to the conversion candidate includes katakana (step 20).
9).

Here, when the input character string includes katakana, the candidate character string generation unit 111 containing kanji and katakana is included.
Searches the heading of the dictionary corresponding to the conversion candidate, and checks whether or not a candidate character string can be created so that the katakana input part becomes kanji or katakana and the hiragana input part becomes hiragana (step 210).

As a result, if it can be created, a candidate character string is created so that the input part of katakana becomes kanji or katakana and the input part of hiragana becomes hiragana,
It is stored in the candidate buffer 115 (step 211).

On the other hand, when the input character string does not include katakana, the kanji-blended candidate character string generation unit 112 searches the dictionary heading corresponding to the conversion candidate so that all the kanji-expressible parts are kanji-expressed. A candidate character string is created and stored in the candidate buffer 115 (step 212). further,
It is inspected whether or not a marker indicating that the headline can be written in Hiragana is attached (step 21).
3) If there is a Chinese character to which a marker is added, a character string in which the Chinese character portion is written in Hiragana is created and stored in the candidate buffer 115 (step 214).

These connection judgments, input character checks,
Each process of creating a candidate character string and storing it in the candidate buffer is performed for all conversion candidates that have been successfully searched for in the dictionary (steps 208 to 215).

When the storage of the conversion candidates in the candidate buffer is completed, the conversion candidate control unit 114 sorts the conversion candidates stored in the candidate buffer 115 according to the number of connection points, the frequency and the learning state in order of priority (step 216). The candidate character strings are displayed in descending order of priority and the selection process is waited for (step 217).

The candidate character string selected by the user is stored in the document buffer 116 by the conversion candidate controller 114 (step 218). The character string stored in the document buffer 116 is displayed on the screen by the display control unit 117 according to the format preset by the format setting unit 104 (step 219). Further, the input notation storage buffer 107, the conversion buffer 109, and the candidate buffer 115 are initialized in preparation for the next input / conversion processing.

In the first configuration of the present invention, step 209, steps 212 to 214, the second step
In the above configuration, the processes of steps 209 to 211 are not performed.

FIG. 4 shows a storage example of each buffer and a dictionary structure when converting an input "tin bell" in an embodiment of the document creation apparatus of the present invention. FIG.
(A) is an input notation. When the input character string includes katakana notation, the notation conversion unit 105 replaces katakana in the character string with hiragana and stores it in the conversion buffer 109 as shown in FIG. As shown in (), the input notation is stored in the input notation storage buffer 107 as it is.

In the kanji / kanji / kanji conversion section 108, a kanji / kanji / kanji conversion dictionary 11 for a character string in which katakana shown in FIG.
A corresponding conversion candidate is searched from 3.

The dictionary headings are stored in a structure as shown in FIG. The first element in each list is the Hiragana notation of the registered word and the heading of the second element. Each element is delimited by "/". For "bellworm" as shown in Fig. (E), "tin" and "bell" and "mushi" and "insect" for each kanji in the heading Readings are associated.

The heading "bell" is provided with a marker "*" indicating that hiragana can be written.
Next, since the input “Suzumushi” includes katakana, the candidate character string including the kanji and katakana candidate character string generation unit 111 creates a candidate character string for the found conversion candidate.

The input "tin bell" includes the katakana notation "pus". So, if you look up the dictionary heading,
Hiragana input part "Suzu" can be written in hiragana, and katakana input part "musi" can be written in kanji, so hiragana input part is hiragana and katakana input part is a candidate for kanji The string "Suzumushi"
Is created. FIG. 6D shows the candidate buffer 115 in which the candidate character strings are stored.

FIG. 5 shows a storage example of each buffer and a dictionary structure when converting the input "Suzumushi" in the embodiment of the document creating apparatus of the present invention. The figure (a) shows the input notation, and the figure (b) shows the conversion buffer 109 storing the input character string. FIG. 4D shows the structure of the dictionary headline, which is exactly the same as FIG. 4E.

A dictionary search is carried out in the kanji / kanji conversion unit 108. In this case, since the input "Suzumushi" does not include katakana, the kanji-blended candidate character string generation unit 112 creates candidate character strings for all the searched conversion candidates.

For the input "Suzumushi", first, a candidate character string "Suzumushi" is created by referring to the Kanji / Kana-Kanji conversion dictionary 113, in which all the Kanji-expressible parts are Kanji-expressed. Next, as shown in FIG. 2 (d), the Chinese character "bell"
Since the marker indicating that the hiragana writing is possible is added to the character string, the candidate character string “Suzumushi” in which the kanji part to which the marker is added is used as the hiragana writing is created. FIG. 7C shows the candidate buffer 115 in which the created candidate character strings are stored.

Next, the dictionary creating apparatus of the present invention will be described. FIG. 6 is a block diagram showing an embodiment of the dictionary creating apparatus of the present invention. The input unit 301 opens the designated dictionary file and reads it into the input buffer 302.
This dictionary file is a normal kanji / kanji conversion dictionary containing kanji and has information on the correspondence between each kanji in the headline and the reading.

The same word detection unit 303 has the input buffer 30.
Sequentially extracting the registered words of the dictionary from 2, and the candidate buffer 30
5, the input buffer 302 is inspected, and if there is a different word identical word of the registered word extracted, it is stored in the candidate buffer 305.

The same word merging unit 304 uses the candidate buffer 30.
When a plurality of registered words are stored in 5, the registered word with the most Kanji writing part is retrieved from the candidate buffer 305,
After a marker is added to a Kanji written in Hiragana with another different notation same word, it is stored in the output buffer 306. The output unit 307 writes the data stored in the output buffer 306 in a dictionary file and outputs the dictionary file.

Next, the operation of the embodiment will be described. FIG. 7 is a flow chart showing the flow of processing from the reading of the input dictionary file to the writing of the output dictionary file in the embodiment of the dictionary creating apparatus of the present invention. First, the input dictionary file designated by the input unit 301 is opened (step 401), and the dictionary data is input to the input buffer 3
02 (step 402).

The same word detection unit 303 has the input buffer 30.
The registered word is fetched from No. 2 and stored in the candidate buffer 305 (step 403). Then, it is searched whether the different-sentence same word of the fetched registered word is stored in the input buffer 302 (step 404). If it is stored, it is fetched from the input buffer 302 and the candidate buffer 305 is fetched. (Step 405).

When the processes of steps 404 and 405 are performed until there are no different words in the input buffer 302,
Of the registered words stored in the candidate buffer 305 by the same word merging unit 304, the registered word t1 with the largest number of Kanji notation parts.
Are retrieved (step 406).

Next, the same word detecting unit 303 checks whether or not a plurality of registered words are stored in the candidate buffer 305 (step 407), and if a plurality of registered words are stored, another candidate expression is changed from the candidate buffer 305. The registered word t2 of is registered (step 408), and a marker is added to the kanji of the registered word t1 in which hiragana is written in the registered word t2 (step 409).

The process of extracting the registered words and adding the marker is repeated for the registered words stored in the candidate buffer 305 (steps 408 to 409). Finally, the registered word t1 is stored in the output buffer 306 (step 41).
1).

As described above, the process of detecting differently-identified same words and the process of merging differently-identified same words are repeated for all registered words stored in the input buffer 302 (steps 403 to 412). The output unit 307 writes the registered word stored in the output buffer 306 to the output dictionary file (step 413),
The dictionary file is closed (step 414).

FIG. 8 shows a state of merging differently-identified same words in an embodiment of the dictionary creating apparatus of the present invention.
FIG. 8A shows the contents of the input buffer, and the structure of the dictionary headings is exactly the same as in FIG.

In the same word detecting unit 303, the differently written same word is searched from the registered words stored in the input buffer 302, and is stored in the candidate buffer 305 as shown in FIG. In the same word merging unit 304, the candidate buffer 3
Of the registered words stored in 05, the registered word t1 having the largest number of Kanji notation parts is taken out, the different notation same word t2 is successively taken out, and they are compared to add a marker.

FIG. 7C shows the registered words t1 and t2 at that time. The Chinese character "bell" of t1 is written in hiragana at t2. As a result, the marker “*” is added to the Chinese character “bell” as shown in FIG.
6 is stored.

[0058]

As described above, according to the present invention, only the word having the most Kanji notation part among the same words having different notations is registered, having the information on the correspondence between each Kanji of the heading and the reading. Using the kana-kanji conversion dictionary containing kanji, the part where the user inputs katakana with the intention of writing kanji is converted to kanji.
It is possible to create and output candidate candidates in which hiragana input is intended for hiragana notation and hiragana is entered.

As a result, the conversion candidates for the notation intended by the user can be obtained without increasing the capacity of the dictionary, and as a result, it is possible to improve the conversion efficiency of a character string containing kanji.

[Brief description of drawings]

FIG. 1 is a block diagram showing the configuration of a document creation device according to an embodiment of the present invention.

FIG. 2 is a flowchart showing a processing flow from character input to conversion candidate selection processing in the embodiment.

FIG. 3 is a flowchart showing a processing flow from character input to conversion candidate selection processing in the embodiment.

FIG. 4 is a diagram showing a storage example of each buffer and a dictionary structure when an input character string includes katakana notation in the same embodiment.

FIG. 5 is a diagram showing a storage example of each buffer and a dictionary structure when the input character string does not include katakana notation in the embodiment.

FIG. 6 is a block diagram showing the configuration of a dictionary creating device according to an embodiment of the present invention.

FIG. 7 is a flowchart showing a processing flow from reading an input dictionary file to writing an output dictionary file in the embodiment.

FIG. 8 is a diagram showing a state of merging differently-identified same words in the embodiment.

[Explanation of symbols]

101 ... Input device, 102 ... Input control unit, 103 ... Input buffer, 104 ... Formatting unit, 105 ... Notation conversion unit,
106 ... Input notation storage unit, 107 ... Input notation storage buffer, 108 ... Kanji mixed kana / Kanji conversion unit, 109 ... Conversion buffer, 110 ... Conversion determination unit, 111 ... Kanji / Katakana mixed candidate character string generation unit, 112 ... Kanji mixed candidate character Column generation unit, 113 ... Kana-kanji conversion dictionary containing kanji, 11
4 ... Conversion candidate control unit, 115 ... Candidate buffer, 116 ...
Document buffer 117 ... Display control unit 118 ... Display buffer 119 ... Display device 301 ... Input unit 302 ...
Input buffer 303 ... Same word detection unit 304 ... Same word merging unit 305 ... Candidate buffer 306 ... Output buffer 307 ... Output unit.

Claims (8)

[Claims] 1. A document creation device for converting kana-kanji character string information into kana-kanji characters and obtaining conversion candidate information thereof, wherein the input notation storage stores the inputted kanji-mixed kana character string information in the notation state. And means for converting the katakana into hiragana when the input character string information stored in this input notation storage means contains hiragana and information about the correspondence between each kanji of the heading and the reading. , Kanji-mixed Kana-Kanji conversion dictionary in which only the words with the most Kanji-description parts are registered in the same word with different notation, and this Kanji-mixed Kana-Kanji conversion dictionary is searched, and the above-mentioned notation conversion means converts Katakana to Hiragana directly. Kana-mixed kana and kanji-mixed kana that obtains conversion candidate information for the selected character string information and determines the connection of the conversion candidate information. Regarding conversion candidate information for which connection determination is permitted by the character conversion means, the katakana part of the input character string information is converted to kanji or katakana or the hiragana part of the input character string information is hiragana from the heading registered in the kanji mixed kanji conversion dictionary. And a candidate character string generation unit for mixing kanji and katakana that creates the candidate character string information of the notation according to the conversion instruction in the input. 2. A kanji-kana kana-kanji conversion method of a document creation device for converting kana-kanji kana-character string information into kana-kanji and obtaining conversion candidate information thereof, which has information on correspondence between each kanji of a heading and reading. , Using the Kana-Kana conversion dictionary in which only the words with the most Kanji notation in the same notation are registered, the Katakana part of the input character string information is converted to Kanji or Katakana, or the input character string information By replacing the hiragana part with hiragana, the candidate character string information of the notation according to the conversion instruction in the input is created, and the created candidate character string information is output as conversion candidate information. Method. 3. A document creation device for converting kana-kanji character string information into kana-kanji characters to obtain conversion candidate information thereof, which has information relating to correspondence between each kanji of a heading and reading and has the same word with different notation. Only the words with the most Kanji notation in the list are registered, and the Kanji-Kana-Kana conversion dictionary in which the Kanji-Kana-Kana-Kanji conversion dictionary in which the Kanji characters that can be written in Hiragana in the registered headings are provided with marker information and this Kanji-Kana-Kana conversion dictionary Then, the conversion candidate information for the input character string information containing mixed kanji is obtained,
Input character string information from the headline registered in the Kanji-Kana Kanji conversion dictionary for the Kanji-Kana Kanji conversion unit that determines the connection of the conversion candidate information and the conversion candidate information for which the connection determination is permitted by the Kanji-Kana Kanji conversion unit. All the parts that can be written in kanji are written in kanji, and further, a candidate character string generation means for mixing kanji that creates candidate character string information in which the part that can be written in hiragana with marker information is written in hiragana is provided. Characterized document creation device. 4. A kana-kana kana-kanji conversion method of a document creation device for converting kana-kanji kana-character string information into kana-kanji and obtaining conversion candidate information thereof, which has information on correspondence between each kanji of a heading and reading. , Among the same words with different notations, only the words that have the most kanji notation parts are registered, and using the kanji-mixed kana-kanji conversion dictionary where marker information is added to the kanji that can be written in hiragana in the registered headings, All the characters that can be written in Kanji in the input character string information are written in Kanji, and further, the candidate character string information in which the Kanji characters that can be written in Hiragana with marker information is written in Hiragana is created. A kana-kanji conversion method that mixes kanji and is characterized by outputting information as conversion candidate information. 5. A document creation device for converting kana-kanji into kana-mixed kana-character string information and obtaining conversion candidate information thereof, wherein the input notation storage stores the inputted kanji-mixed kana-character string information in the notation state. And means for converting the katakana into hiragana when the input character string information stored in this input notation storage means contains hiragana and information about the correspondence between each kanji of the heading and the reading. , Only the words with the most Kanji notation parts are registered in the same word with different notation, and the Kanji conversion Kana-Kanji conversion dictionary in which marker information is added to the Kanji that can be written in Hiragana in the registered heading Search the mixed kana-kanji conversion dictionary, and input candidate character string information or conversion candidate information for the character string information in which katakana is converted to hiragana by the above notation converting means. If the input character string information contains Katakana, the conversion candidate information for which the connection determination of the conversion candidate information is determined, and the conversion candidate information for which the connection determination is permitted by the Kanji mixing kana Kanji conversion means Candidate character strings that correspond to the conversion instruction in input by replacing the katakana part of the input character string information with kanji or katakana from the heading registered in the mixed kana-kanji conversion dictionary, or by replacing the hiragana part of the input character string information with hiragana The Kanji-Kana-Kana Kanji conversion is performed on the Kanji-Kana-Kana Kana-Kana Kanji conversion unit that creates the information and the conversion-Kanji-Kana-Kana Kanji conversion that the connection determination is permitted by the Kanji-Kana Kana-Kana conversion unit when the input character string information does not include Katakana. Kanji notation of input character string information is possible from the heading registered in the dictionary All the Kana characters are written in Kanji, and the candidate character string generation means that creates the candidate character string information in which the Hiragana notation that allows the writing of Hiragana with marker information is created and the input character string information A document creation apparatus, comprising: a conversion determination unit that determines whether to use the Kanji / Kana / Kana character mixture candidate character string unit or the Kanji / Kanji character mixture candidate character string generation unit depending on whether or not Katakana is included. 6. A kana-kana kana-kanji conversion method of a document creation device for converting kana-kanji kana-character string information into kana-kanji and obtaining conversion candidate information thereof, which has information on correspondence between each kanji of a heading and reading. , Among the same words with different notations, only the words with the most Kanji notation parts are registered, and marker information is added to the Hiragana notationable Kanji of the registered heading. It is checked whether or not the character string information includes katakana. If the input character string information includes katakana, the katakana part of the input character string information is converted to kanji or By using katakana or by replacing the hiragana part of the input character string information with hiragana, the candidate character string for the notation according to the conversion instruction in input When information is created and if the input character string information does not include katakana, all the kanji notation of the input character string information from the heading registered in the Kanji-mixed kana-kanji conversion dictionary is set to kanji and the marker A kana-kanji kana-kanji conversion method characterized by creating candidate character string information in which hiragana-expressible kanji parts provided with information are written in hiragana and outputting the created candidate character string information as conversion candidates. 7. A kanji-kana kana-kanji conversion dictionary having information on the correspondence between each kanji of a heading and a phonetic reading, a kanji-kana kana-kanji conversion dictionary for detecting the same word with different notations, and the same word Among the same words with different notations detected by the detection means, the most common word in kanji is given marker information to the kanji written in hiragana in other notations, and the rest is deleted. A dictionary creating device comprising: a means for creating a kanji conversion dictionary containing kanji mixed with only the word having the most kanji notation part among the same words having different notations. 8. A kanji-kana-kanji conversion dictionary having information on the correspondence between each kanji of a heading and its reading, if there are multiple words registered for each heading with a different notation, one of the registered headings Marker information is added to the kanji written in Hiragana, and only the word with the most kanji is left, and other words with different kanji are deleted. A dictionary creation method characterized by creating a kana-kanji conversion dictionary in which only kanji is registered.