JPH0154742B2 - - Google Patents

Description

[Detailed Description of the Invention] <Technical Field> The present invention relates to a kana-kanji conversion device that converts kana characters or kana sentences (or Roman characters) into corresponding kanji or kanji-mixed sentences according to a kana-kanji conversion dictionary. The present invention relates to a kana-kanji conversion device having a function of converting an abbreviation kana sentence into a kanji or a kanji-mixed sentence corresponding to the abbreviation in response to input of the abbreviation kana sentence.

<Prior Art> Recently, Japanese document processing devices (Japanese word processors) have been put into practical use for the purpose of improving document processing efficiency. This Japanese document processing device inputs a kana sentence using an input device such as a kana keyboard, searches a kana-kanji conversion dictionary, and outputs the corresponding kanji or a sentence mixed with kanji. In such devices, when creating documents, abbreviations such as "Japan National Railways" or "flukan" are sometimes used. In other words, their official names are ``National Railway'' and ``Epidemic Flu.''

In conventional devices, it is common to register formal names in a kana-kanji conversion dictionary, and in order to convert the abbreviations as well, it is necessary to register all abbreviations in the kana-kanji conversion dictionary. The problem with registering abbreviations in this dictionary is that the memory capacity of the kana-kanji conversion dictionary becomes enormous.

Conventionally, therefore, conversion processing is performed for each single kanji character, such as ``koku'' and ``tetsu,'' to avoid increasing the capacity of the dictionary. However, performing the conversion process for each single kanji character in this way is extremely difficult to operate, and is not suitable for the current situation where high-speed processing is required.

<Purpose> In view of the above-mentioned problems, the present invention enables kana-to-kanji conversion of abbreviations without increasing the capacity of the kana-to-kanji conversion dictionary, thereby improving the operability of the conversion process. It is.

<Embodiment> FIG. 1 is a block diagram showing an outline of a kana-kanji conversion device (word processor) that implements the abbreviation conversion method of the present invention. In FIG. 1, reference numeral 1 denotes an editing control unit (CPU), which is controlled by a program stored in RAM3. A program is previously loaded and held in this RAM 3 from an external memory means (not shown).

Reference numeral 2 denotes a keyboard device equipped with character keys such as alphabets and kana, and various instruction keys for editing control.

4 is a RAM for holding various data, including a text buffer 41 for storing edited text data, a kana buffer 42 for introducing and storing kana character strings input from the keyboard device 2, and other buffers and flags. It is made up of an area 43 that includes, etc.

Reference numeral 5 denotes a dictionary memory control unit for controlling the dictionary memory 6, and it mainly comprises a result buffer 51 for storing candidate data read from the dictionary memory 6, and a plurality of address circuit units 52 for addressing the dictionary memory 6. Teru.

7 is a display control section of the CRT display device 8, and 9 is a printer control section of the dot printer 10.

FIG. 2 shows the configuration of the dictionary memory 6 described above,
Area 61 contains kana entry words, area 62 contains kanji or kanji-mixed sentences (converted words) corresponding to the kana entry words, area 63 contains part-of-speech information, and area 64 contains readings corresponding to each single kanji. In the area of 65 kana, phrases that can be connected are stored respectively.

Specifically, each character code string is stored in the areas 61 and 62, and the address value of the corresponding dictionary is stored in the area 65.

3 and 4 are flowcharts showing the operation of the configuration shown in FIG. 1, and the operation of the configuration shown in FIG. 1 will be explained below in accordance with this flowchart.

First, to explain the abbreviation "Japan Railways" as an example, the kana character string "Kokutetsu" is input using the kana character keys of the keyboard device 2. Then, RAM4
Kana characters are introduced and stored in the kana buffer 42 each time a kana is input.

Subsequently, when a conversion is instructed, this kana character string is supplied to the dictionary memory control section 5 and the dictionary memory 6 is searched.

In other words, the headword in the dictionary memory 6 is "Kokutetsu".
A search is performed to see if there is a headword.

In this case, the headword cannot be found because it is an abbreviation of "Kokutetsu," but if it is a regular kana character string that is not an abbreviation, the headword will be detected and the corresponding kanji character string will be detected. The sentence is read into the result buffer 51 and displayed on the CRT display device 8.

If the displayed sentence mixed with kanji is the target conversion sentence, conversion is determined by a selection operation and is stored in the sentence buffer 41. If there are multiple homophones, the next candidate sentences containing kanji are sequentially read out, and the desired converted sentence is selected.

However, since the headword "kokutetsu" is an abbreviation and is not detected, processing is performed as if there were no candidates (steps 100 to 102 in FIG. 3).

Next, if there are no candidates in the dictionary memory 6, then the character string in the kana buffer 42 is divided into front and back parts. In this division, first, it is divided into 1/2, and the headwords containing the kana characters are searched for the former and latter kana character strings, respectively.

In other words, in this example, it is divided into "koku" and "tetsu", and the kana character strings before and after these are first supplied to the dictionary memory control unit 5 to search for a headword that includes this kana character string, and when the headword is detected. Furthermore, a search is made to see if this entry word contains a subsequent kana character string.

Through this process, it is determined whether there is an official name that includes the target abbreviation, and in this case, the headword "Kokuyuutetsudo" is detected for "Kokutetsu".

Subsequently, individual single kanji characters corresponding to the headword "kokutetsu" are extracted from the converted words corresponding to the detected headword and displayed on the CRT display device 8. If this is the desired abbreviation, it is selected and stored in the text buffer 41 as described above (steps 103-108 in FIG. 3).

In addition, when dividing the character string of the above-mentioned kana buffer 42 into front and rear parts, the dictionary memory 6 was initially divided into 1/2 parts and the former part and the latter part were searched, but at this time, if the entry word could not be searched. To do this, add one kana character to the previous kana character string, divide it into "kokute" and "tsu", or divide it into "ko" and "kutetsu", etc. This allows the dictionary memory 6 to be searched.

Next, the abbreviation conversion process will be explained in more detail with reference to the flowchart of FIG.

The kana character string "Kokutetsu" of the above-mentioned kana buffer 42 is divided into "Koku" and "Tetsu" in the first and second stages, and this first stage kana character string "Koku" is supplied to the dictionary memory control unit 5, The dictionary memory 6 is searched for headwords containing this kana character string.

In this case, if the dictionary memory 6 has the configuration shown in FIG. 2, the first entry word "Kokuyuutetsudo" having the entry word "Koku" is detected, and along with the converted word corresponding to this entry word, 63 , 64 and 65 are read out to the result buffer 51 (steps 200-202 in FIG. 4).

Next, the detected headword and the number of pronunciation kana characters for the kanji (64 data) are matched to confirm whether the headword for "koku" can be divided into one or more kanji.

Now, “koku” has two letters, and the pronunciation of the kanji “kuni” that corresponds to “koku” in the headword detected in the dictionary (kokuyuutetsudo) is also two letters, so for “koku” Since it is a single kanji, it is determined that it can be divided into kanji. In other words, the position where the above-mentioned first and second stages are divided (divided into 1/2) is determined to be a position that can be divided as a kanji (step
203).

Next, it is determined whether the flag is ON or not, and since the entry word corresponding to "koku" has been detected, the flag is in the OFF state. Therefore, the process proceeds to the next step and uses the converted word corresponding to this detected entry word. It is determined whether the kanji string includes a delimiter code and can be used as an abbreviation (steps 204 and 205).

That is, in the dictionary memory 6 shown in FIG. 2, for the registered kanji that can be used as an abbreviation, a delimiter code (indicated by a slash "/" in FIG. 2) is placed at the position where the kanji is divided. It has been inserted.

For example, the kanji string corresponding to ``kokuyuutetsudo'' is registered as ``nationally owned/railway,'' indicating that the kanji string can be used as an abbreviation.

Therefore, in this case, it is determined that the delimiter code exists, and the process proceeds to the next step to store the latter kana character string "tetsu" in the dictionary memory control unit 5.
and then checks whether this “tetsu” is included in the detected headwords (step
206, 207).

Next, it is checked whether this "tetsu" is split into one or more kanji, and now "tetsu" is two characters, and the dictionary's detected entry word (kokuyuutetsudo) is "tetsu". Since the reading kana for the kanji character ``tetsu'' corresponding to ``tetsu'' is also two characters and is one kanji character for ``tetsu'', it is determined that it can be divided into kanji characters (step 208).

Next, the kanji corresponding to the above-mentioned ``koku'' and ``tetsu'' are extracted from the kanji string of the converted word ``nationally owned/railway'' to obtain the kanji strings of the abbreviated words ``koku'' and ``tetsu.'' (Step 209) ).

In this way, for the kana character string of an abbreviation word, as mentioned above, the kana character string of the abbreviation word is divided into front and rear parts, and the headword containing the divided kana character strings in the first and second part is stored in the dictionary memory. Then, from the converted word corresponding to the searched entry word, the kanji corresponding to each of the divided kana character strings are read from the dictionary memory to obtain a converted word corresponding to the abbreviation word.

Next, as another example, a case where the abbreviation "Riyuukan" is input will be described.

First, the kana character string input to the kana buffer 42 is supplied to the dictionary memory control section 5 and the dictionary memory 6 is searched.

The dictionary is searched for this headword "Riyukan", but as shown in FIG. 2, no kanji with this abbreviation as a headword is registered, so it is processed as having no candidates.

Next, the character string in the kana buffer 42 is divided into two parts. In other words, it is divided into "Riyuu" and "Kan". The previous kana character string "Riyuu" is then supplied to the dictionary memory control unit 5, which searches the dictionary memory 6 for the headword "Riyuu". ``Ryuuki'' is selected. Subsequently, after it is determined that this ``riyu'' can be divided into one kanji, it is checked whether the kanji string contains a delimiter code, that is, whether it can be used as an abbreviation. However, in this case, since there is no delimiter code, the dictionary is searched again for the next candidate.

As a result, the address m+1 of the dictionary in FIG.
A search is made for the entry words .about.m+3, but since none of them can be used as an abbreviation, a flag is turned on to indicate that there are no candidates for the search for this abbreviation (steps 202→210).

When this flag is turned ON, the dictionary will be searched first again.

That is, this time it is checked whether the kanji provided in the dictionary in advance has a connected idiom.

This divides the kana character string "Riyuukan" into the first and second parts, and supplies the first part "Riyuu" to the dictionary memory control section 5, and searches for a headword containing this "Riyuu" (Step 201).

After that, the process proceeds to steps 202, 203, and 204, and in step 204, the flag is set as described above.
Since it is in the ON state, the process advances to step 211 and it is checked whether there are any connectable idioms.

In this example, the first headword is m in Figure 2.
The address "Riyuuki" is selected, and since there is a connectable compound word, the position of the headword of this compound word is addressed, and the latter kana character string "kan" is included in the headword of "Gensho". is checked. In this case, there is no "kan", so the dictionary is searched again for the next candidate.
In this example, for "Riyuu" in "Riyuu" and "Kan", the headword "Riyuukosei" which is the m+2 address of the dictionary is selected, and the position of the headword "Kanbou" is selected as the connected idiom. When ``kan'' is addressed, it is determined that ``kan'' is included.

This causes the process to proceed to step 214, where it is checked whether the part corresponding to this "kan" can be divided into one or more kanji. In this case, "kan" becomes the kanji "kan", and it is determined that it can be divided into kanji units, and the process proceeds to step 209.

And here, the entry word “Riyuukosei”
The kanji ``flow'' corresponds to ``riyuu,'' and the kanji ``kan'' corresponds to ``kan'' in the headword ``kanbou.''
The kanji ``flukan'' is obtained by cutting out the word ``flukan''.

In this case, the kanji corresponding to the abbreviation are selected from the two converted sentences in the dictionary and converted.

The above flag will be turned OFF once the conversion is complete.
become.

<Effects> As described above, in the present invention, the dictionary memory is equipped with additional information indicating kanji or compound words that can be connected to the converted word corresponding to the headword in relation to the converted word, For an input kana character string of an abbreviation word for which there is no conversion word in the dictionary memory that uses the input kana character string as a headword, this input kana character string is divided into at least the first stage and the second stage, and the kana characters of the divided first stage are processed. Search the dictionary memory for the headword containing the column, and use the conversion word corresponding to the headword detected by the search and the additional information to determine whether the kanji corresponding to the divided kana character string in the subsequent stage is included. Based on this determination, when it is detected that the kana character string in the latter part contains a corresponding kanji character, the kanji corresponding to the divided former and latter kana character strings are read from the dictionary memory and the abbreviation is read out. It is designed to obtain the converted word corresponding to
The dictionary memory described above does not register abbreviations in correspondence with headwords, but instead stores them as additional information in relation to converted words. This avoids an increase in the memory capacity, and moreover, the abbreviations can be stored in Since it does not convert each kanji character by character, it is easy to operate, and abbreviations can be efficiently converted without increasing the size or cost of the device.

[Brief explanation of drawings]

FIG. 1 is a block diagram showing the configuration of a kana-kanji conversion device implementing the abbreviation conversion method of the present invention, FIG. 2 is a diagram showing the configuration of a dictionary memory, and FIGS. 3 and 4 are the device shown in FIG. 1. This is a flowchart showing the operation of the kana-kanji conversion process. 1: Edit control unit (CPU), 2: Keyboard device, 4: RAM, 5: Dictionary memory control unit, 6: Dictionary memory, 8: CRT display device, 41: Text buffer, 42: Kana buffer, 51: Result buffer .

Claims (1)

[Scope of Claims] 1. A dictionary memory storing a kana headword and a kanji or kanji-mixed sentence (hereinafter referred to as a converted word) corresponding to the headword, and a predetermined kana character string inputted from a kana character input means. In the kana-kanji conversion device which obtains a converted word corresponding to the inputted kana character string by reading a converted word having the relevant kana character string as a headword from the dictionary memory, the dictionary memory stores the corresponding converted word related to this converted word. Equipped with additional information indicating kanji or compound words that can be connected to the conversion word, this input kana string is used for input kana strings of abbreviations without conversion words that use the input kana string as the headword. The dictionary memory is searched for a headword containing the divided kana character string in the first stage and the second stage, and the converted word corresponding to the headword detected by the search and the above-mentioned addition are processed. It is determined from the information whether the kanji corresponding to the kana character string in the latter part of the division is included, and based on the determination that it includes the kanji character corresponding to the kana character string in the latter part, the first part of the division is determined. An abbreviation word conversion method in a kana-kanji conversion device, characterized in that a conversion word corresponding to the abbreviation word is obtained by reading out the kanji characters corresponding to each subsequent kana character string from a dictionary memory.