JP2634926B2 - Kana-Kanji conversion device - Google Patents

Description

The present invention is mainly used in an information processing apparatus having a kana-kanji conversion dictionary, such as a Japanese word processor, and converts an input kana character string to a kanji kana character. The present invention relates to a kana-kanji conversion device for converting kana to kanji.

(B) Conventional technology In general, in a Japanese word, there are cases where there are a plurality of expressions that are allowed depending on how to attach a sentence. For example, in the case of the word “remember”, there are sendings such as “memorandum”, “memorandum”, “memorandum”, and “memorandum”. In addition, the notation is also used for "remembering", "oboegaki", "remembering", "remembering"
And so on.

In the conventional Kana-Kanji conversion device, in order to input a word with many types of feeds in a desired way, all words with different feeds are registered in the Kana-Kanji conversion dictionary, and the desired word is selected from the dictionary. I do.

Only one kana-kanji notation for one word is registered in the kana-kanji conversion dictionary, and after kana-kanji conversion, the desired kana-kanji conversion is corrected.

The algorithm deletes the sentence. In other words, only the notation in which all of the sending letters are added is registered in the dictionary, and the first notation is used as the first candidate for conversion, and the sending letters are sequentially deleted from there. For example, in the case where there are four candidates for the sending word, such as "memorandum", "memorandum", "memorandum", and "memorandum", as in the above-mentioned "notes", the notation with the largest number of characters in the sending word Is the first candidate, and the first candidate (memorandum) → the second candidate (memorandum or memorandum) → the third candidate (memorandum) is automatically deleted.

Such a method is used.

However, as described above, registering all words having different sending words in the kana-kanji conversion dictionary causes a decrease in the number of registered words within a limited dictionary capacity.

Also, as described above, when only one notation is registered for one word, each time the word is read out, the notation is corrected (addition or deletion of the notation).
And the correction is performed after the conversion is determined, so that the learning function is not effective.

Furthermore, as described above, in the case where a feed word is deleted by the algorithm, it is difficult to perform notation processing in a direction in which the feed word is added. For example, in the case of “remember”, the first candidate “memorandum” In this case, there is a problem that it is difficult to set the next candidate as “memorandum” (additional processing for sending).

Therefore, in order to solve such a problem, the present inventors store the pattern of the transmission in a table separately from the kana-kanji conversion dictionary, and change the transmission based on the table to change the transmission. We applied for a Kana-Kanji conversion device that did not require correction operations such as addition and deletion (Japanese Patent Application No. 1-133737).

(C) Problems to be Solved by the Invention However, in the kana-kanji conversion device, for example, if the word “oboegaki”
Although it can be converted to "memo,""memo,""memo,""oboegaki," etc., it cannot be converted to "memo," or "memo." Did not. That is, for example, if the word is "scoop ruler", only a part of the multiple kanji notation is flattened, such as "shakushi ruler" or "squeeze down" or "squeeze down". It could not be converted to kana notation.

The present invention has been made in view of such circumstances, and for a word having a plurality of notations in a feed,
The kanji / hiragana notation pattern is also stored for words that have multiple kanji notation characters and a part of which has been changed to hiragana, as well as storing the sending pattern as a table separately from the kana-kanji conversion dictionary. Is stored as a table separately from the Kana-Kanji conversion dictionary, and when those words are read out, the spelling and spelling are changed based on each table, and the spelling and spelling correction operation as in the conventional method is performed. The present invention provides a kana-kanji conversion device which eliminates the need for.

(D) Means for Solving the Problems FIG. 1 is a block diagram showing the configuration of the present invention.

As shown in the figure, the present invention provides a dictionary means 101 which stores a large number of words including kanji together with their reading information, and stores a word having fluctuation in notation by attaching an identification code 102 thereto.
And the identification code 102, a plurality of types of bit patterns indicating the presence or absence of a kana for each character position, and a kana storage means 103 storing a table, and a plurality of types of kanji / hiragana for each character position. Kanji / hiragana notation storage means 104 storing bit patterns representing notations as a table
Reading means 106 for reading a word corresponding to the reading information input from the input means 105 from the dictionary means 101; and a word read by the reading means 106,
When 02 is added, the notation of the word is sequentially converted to the plurality of types of notation in the form of plural words based on the table stored in the storage device 103, and the kanji / hiragana is changed according to the identification code. Conversion means 107 for converting a part or all of a plurality of kanjis of the word into plain notation based on a table stored in the notation storage means 104;
A kana-kanji conversion device comprising a display unit 108 for displaying a word whose notation has been converted by the conversion unit 107.

Note that the dictionary means 101 in the present invention may be any one that can store a large number of words including kanji together with its reading information, and can store words with fluctuations in notation with identification codes. Alternatively, an external storage medium such as a floppy disk device or a magnetic disk device is used.

As the identification code 102, various codes such as identifiable numbers and symbols are used.

Further, as the sending storage unit 103 and the kanji / hiragana notation storage unit 104, any one can be used as long as it can store a plurality of types of bit patterns representing kanji / hiragana notation for each character position as a table. RO as in means 101
M or an external storage medium such as a floppy disk device or a magnetic disk device is used.

As the input means 105, any means can be used as long as it can input flat character strings and the like as kana and kanji reading information, and for example, a keyboard, a tablet device, an OCR, a magnetic tape device, or the like is used.

Further, as reading means 106, any means can be used as long as it can read words corresponding to the reading information input from input means 105 from dictionary means 101, and conversion means 107
When the identification code 102 is attached to the word read by the reading means 106, the sending storage means 1
The notation of the word is sequentially converted into a plurality of types of kana-kana notation based on the table stored in 03, and based on the table stored in the kanji / hiragana notation storage means 104 according to the identification code. Anything that can convert some or all of the multiple kanji of the word into plain notation,
In general, it is convenient to use a microprocessor as the reading means 106 and the converting means 107.

As the display unit 108, any device capable of displaying the word converted by the conversion unit 107 may be used, and a printer, a display device, a magnetic tape device, a magnetic disk device, or the like is used. It is preferable to use a display device.

(E) Operation As shown in FIG. 1, according to the present invention, the input means 10
When reading information is input from 5, a word corresponding to the reading information is read from the dictionary means 101 by the reading means 106.

At this time, if the identification code 102 is attached to the word to be read out, the conversion unit 107 converts the word notation into a plurality of types of notation based on the table stored in the no-feed storage unit 103. The kanji / hiragana notation storage means 10 is sequentially converted and according to the identification code.
Based on the table stored in 4, some or all of the kanjis of the word are converted to a plain representation and displayed by the display means 108.

Therefore, after the word kanji is converted based on a preset table, part or all of the kanji of the word is converted to hiragana. Since the registration of words having different sending patterns and notation patterns can be performed by one word, the dictionary capacity can be saved.

(F) Embodiment Hereinafter, the present invention will be described in detail based on an embodiment shown in the drawings. Note that the present invention is not limited to this.

FIG. 2 is a block diagram showing an embodiment of the present invention.

In this figure, reference numeral 1 denotes a keyboard provided with a kana key, a function key, a kana-kanji conversion key for instructing kana-kanji conversion, or an execution key for confirming the input content, and controls word reading information by kana. Input to section 2.

The control unit 2 includes a microprocessor, and performs various data processing described below according to a control program written in a program memory 3 formed of a ROM.

Reference numeral 4 denotes an input buffer for storing word reading information based on kana input from the keyboard 1.

Reference numeral 5 denotes a kana-kanji conversion dictionary composed of a ROM, which is an independent word dictionary 6 storing a large number of words including kanji together with their reading information, a part-of-speech table 7 storing part-of-speech information, a transmission part table 8 storing transmission-side information, Has a kanji / hiragana notation table 9 which stores information for writing kanji in hiragana.

Of the words stored in the independent word dictionary 6, words having a plurality of types of kanji (hereinafter referred to as words with fluctuations) and words having a plurality of types of kanji / hiragana expressions (hereinafter, fluctuations in notation) Are referred to as “words”) with a notation pattern number as described later.

In addition, in the sending table 8, a sending pattern for a word having a written pattern number in the independent word dictionary 6 is stored as a table.
In the hiragana notation table 9, a kanji / hiragana notation pattern for converting a part or all of a plurality of kanjis of a word having a notation pattern number into a hiragana notation is stored as a table. It is remembered.

Reference numeral 10 denotes a display device including a CRT display, an LC (liquid crystal) display, an EL display, and the like.

The control unit 2 stores the reading information of the word input from the keyboard 1, that is, the kana character string in the input buffer 4,
When the kana-kanji conversion is instructed for the kana character string, the word corresponding to the kana character string is read from the independent word dictionary 6.

When a notation pattern number is added to the read word, the notation of the word is represented by a plurality of kinds of notation based on the notation pattern located at the pointer indicated by the notation pattern number in the notation table 8. And a part or all of the kanji of the word is converted to hiragana based on the notation pattern located at the pointer indicated by the kanji / hiragana notation number in the kanji / hiragana notation table 9. And displays it on the display device 9.

FIG. 3 is an explanatory diagram showing a storage format of the independent word dictionary 6.

As shown in the figure, the independent word dictionary 6 stores, for each word, a word length of 10, a duplication of 11, a pronunciation of 12, a part of speech number of 13, and a notation of 14.
The item of notation pattern number 15 is provided only for words having fluctuations in notation and notation.

The word length 10 is the record length for one word, the duplication 11 is the information that only the changed part of the same notation is stored to avoid the storage of the duplicated part of the word, and the reading 12 is , And a plain notation as a heading are stored.

The part-of-speech number 13 is a part in which part-of-speech information such as a noun and a verb is stored. The part-of-speech information for each word is separately stored in the part-of-speech table 7.
Stores a pointer (numerical value) indicating in which part of the record of the part of speech table 7 the part of speech information of the word is stored. The notation 14 stores the notation of a word including a kanji.

These have a word length of 10, a duplication of 11, a reading, and a part of speech number of 1.
3. The notation 14 conforms to the structure of a conventionally known independent word dictionary.

In the notation pattern number 15, similarly to the part of speech table 7 described above, a pointer (numerical value) indicating in which record of the kana character table 8 the word kana pattern and the kanji / hiragana notation pattern are stored. Is stored.

Whether or not the word has the notation pattern number 15 is determined based on whether or not the part of speech number 13 exceeds a certain numerical value. That is, part-of-speech information on a word having fluctuations in the sentence notation is collectively registered in the second half of the part-of-speech table 7, and by checking the part-of-speech number 13, it is determined whether or not the notation pattern number 15 is added to the word. It is possible to judge.

FIG. 4 is an explanatory diagram showing a storage format of the sending table 8.

As shown in the drawing, in the sending table 8, the sending pattern and the kanji / hiragana notation pattern are bit patterns in the order of the notation pattern numbers, and the first candidate, the second candidate,
The third candidate,... Are sequentially stored. For example, if the notation pattern number is n, the first candidate (000100
00), the second candidate (01010000), and the third candidate (00000000).

In the contents of this sending table 8, the sending part pattern portion is made to correspond to the kanji notation in byte units. For example, in the case of a word having a spelling pattern of the notation pattern number n in FIG. 4 called "remembering", the read character data is (Kaku = 2) (E = 1) (Book = 1) (K = 1) = 2111, and this four-digit notation, as shown in FIG.
One bit is allocated from the beginning. Then, as shown in FIG. 6, the notation in which the bit is "1 (on)" is deleted.

With such a processing method, if the sentence pattern of the word selected once is stored in the learning buffer, the desired sentence of the word can be expressed by the first conversion from the second time. Can be obtained.

FIG. 4 is a table in which the feed pattern and the kanji / hiragana notation pattern are tabulated as 1-byte bit pattern information.

FIG. 7 is an explanatory diagram showing the contents of the sending table 8 shown in FIG.

As shown in the figure, in the case of the above-mentioned “remembering”, which is the feed pattern indicated by the notation pattern number n in FIG. 4, the first candidate (memorandum) → the second candidate (memorandum) → The third candidate (memorandum) → the fourth candidate (memorandum) → the fifth candidate (remember) → the sixth candidate (oboegaki) are sequentially converted.

As described above, the bit pattern of the transmission is
For example, in the case of “remember”, the first four bits are registered as bits with information of “1 (on)” and “0 (off)”, and “remember” is input. When converting the kana to kanji characters, the notation character whose feed pattern bit pattern corresponding to the reading data of the first four characters is "1 (on)" is deleted.

Of these transformations, for the fifth and sixth candidates,
Since the first character of a word is not usually deleted, the first character is used as “1 (on)”, that is,
If it is a bit of deletion, all the readings of the word are written in hiragana or katakana.

For the seventh and subsequent candidates, the first bit is “1”.
In the case of “(ON)”, the bits that can be expressed other than representing hiragana or katakana are 6 bits as indicated by the crosses in the figure.
This is used when the first bit is “1 (ON)” and one of the 6 bits from the second bit to the seventh bit is “1 (ON)”. Regards it as a kanji / hiragana notation number and refers to the kanji / hiragana notation table 9 shown in FIG.

That is, the kanji / hiragana notation number serves as a pointer to the kanji / hiragana notation table 9, and this pointer refers to the kanji / hiragana notation table 9.

The kanji / hiragana notation table 9 is stored in the form of an 8-bit pattern in the order of kanji / hiragana notation numbers. That is, for example, when the word is “squeeze down”, it is registered in the independent word dictionary 6 with a notation pattern of “narrow down”, and the read character data is (squeeze = 2) (ri = 1) (inclusive = 1) (M = 1) = 2111

When only the stem is registered, there is a case where there is no last "mu" information.

Therefore, when the character string “SHIBORIKOMU” is converted to Kana-Kanji characters, the sending table 8 as shown in FIG. 9 is searched from the notation pattern numbers 15 added to the words in the independent word dictionary 6, and The first candidate and the second candidate are sequentially converted to Kana-Kanji characters, and in the case of the fourth candidate, the bit pattern of the fourth candidate is converted back to a binary number in a form in which the first bit and the last bit are removed, and A kanji / hiragana notation table 9 as shown in FIG. 10 is referred to. Here, the kanji / hiragana notation number of the fourth candidate in the sending table 8 is the first binary code, and the kanji /
The first notation pattern in the hiragana notation table 9 is “1000
000 ", the first kanji in the kanji notation is changed to a hiragana notation.
Since the first kanji reading is found to be the first and second reading characters, "squeezing" is changed to "shiro" and converted to the notation "squeezing in" as shown in Fig. 10. .

Similarly, in the case of the fifth candidate, the kanji / hiragana notation number of the fifth candidate is the binary code No. 2, and the second notation pattern of the kanji / hiragana notation table 9 is “001000”.
Since it is "00", "in" is changed to "ko" and converted to the notation "to narrow down".

Thus, for example, in the case of the word “Shakushi Jyogi”, the read character data is (ladder = 3) (child = 1) (constant = 3) (rule = 1) = 3131. If the kanji / hiragana notation number is set so as to correspond to the n-th notation pattern in the kanji / hiragana notation table 9 shown in FIG. 10, it can be converted to the notation "shakushi ruler". Can be.

If there are no more than 255 types of kanji and kanji / hiragana notation patterns, the notation pattern number is 1-byte information, the bit pattern information is fixed length, and If the type is 256 or more, the notation pattern number is 2-byte address information, and the bit pattern information is variable-length data by an integral multiple of 1 byte.

The one-byte information of the bit pattern shown in FIG. 7 is an example. For example, in the case of “memorandum”, the first four bits are “1 (on)” and “0”.
(Off) "may be reversed, and FIG. 11 shows an example in that case.

Next, the contents of the processing operation of the control unit 2 will be described with reference to the flowcharts shown in FIGS.

FIG. 12 is a flowchart showing the contents of processing when kana-kanji conversion is performed.

When the kana-kanji conversion is performed, first, a character string is input from the keyboard 1 (step 201), and the character string is stored in the input buffer 4, and when the kana-kanji conversion is instructed (step 202), control is performed. The unit 2 determines whether there is a homonym (step 203). If there is a homonym, candidates for the homonym are extracted and displayed on the display device 9 (step 204). A selection is made (step 205).

Next, the part-of-speech number 13 of the independent word dictionary 6 is searched to determine whether there are fluctuations in the feed and notation (step 206). If there are fluctuations in the feed and the notation, the fluctuations in the feed and the notation are detected. Display device 9 extracted
Are displayed (step 207), and these are selected (step 208).

If the character string is input again, step 20
Returning to 1, if the character string is not input, the kana-kanji conversion processing ends (step 209).

FIG. 13 is a detailed flowchart in the case where the feed length and the fluctuation of the notation shown in FIG. 12 are extracted.

In the case where the infeed and the fluctuation of the notation are extracted, first, it is determined whether or not the notation pattern number is added in the independent word dictionary 6 (step 301). , The sending table 8 is referred to based on the written pattern number (step
302) It is determined whether or not the bit pattern of the one-byte information “Shigana” is a kanji / hiragana notation number (step 303). Then, the information of the corresponding sending pattern is searched (step 304).

Then, one of the sending patterns corresponding to the kanji notation
A feed is created and output from the byte information (step 305).

Also, if the kanji / hiragana is a spelling number, the kanji /
Information of the corresponding notation pattern is retrieved from the hiragana notation table 9 (step 306), and a kanji / hiragana notation pattern is created and output from the notation pattern information and the read character data (step 307). .

In this way, only the words to which the notation pattern numbers are added to the independent word dictionary 6 are registered in advance in the sending table 8 and the kanji / hiragana notation table 9 with the sending pattern and the kanji / hiragana notation pattern, respectively. In addition, by converting the word sentence notation based on those tables, the word can be written in a desired form by a simple conversion operation without increasing the number of registered words in the independent word dictionary 6. It becomes possible.

(G) Effects of the Invention According to the present invention, for a word having a plurality of types of patterns in sentence notation, the pattern is stored as a table separately from the dictionary means, and when the word is read, the pattern is stored. Based on the table, we convert the word notation into multiple types of patterns,
This eliminates the need to add or delete traditional spelling words or convert kanji to hiragana, and saves the dictionary capacity because only one word can be registered with multiple types of spelling patterns. Can be saved.

In particular, for a word having a plurality of types of patterns in the notation, the kanji / hiragana notation pattern of the word is stored as a table separately from the dictionary means. It is possible to easily change the notation of converting to kana. For example, "squeeze in"
In the case of, some or all of a plurality of kanji characters are used, such as a first candidate (refine) → a second candidate (refine) → a third candidate (refine) → a fourth candidate (refine). It can be easily converted to flat.

In addition, for sending, the candidate ranking of the notation,
For example, in the case of "remembering", arbitrarily set, for example, a first candidate (memorandum) → a second candidate (memorandum) → a third candidate (memorandum) → a fourth candidate (memorandum). Therefore, it is possible to perform not only the process of deleting a sentence but also the process of adding and deleting a sentence for each word, in which the conventional algorithm deletes the sentence.

[Brief description of the drawings]

FIG. 1 is a block diagram showing a configuration of the present invention, FIG. 2 is a block diagram showing an embodiment of the present invention, FIG. 3 is an explanatory diagram showing a storage format of an independent word dictionary, and FIG. FIG. 5 is an explanatory diagram showing an example of bit allocation, FIG. 6 is an explanatory diagram showing an example of deleting bits of a sending tag, FIG. 7 is an explanatory diagram showing contents of a sending tag table, FIG. FIG. 8 is an explanatory diagram showing a storage format of a kanji / hiragana notation table, FIG. 9 is an explanatory diagram showing another example of contents of a sending kanji table, and FIG. 10 is a specific example of a kanji / hiragana notation table. FIG. 11 is a diagram corresponding to FIG. 7 showing the case where the bit pattern is reversed, and FIGS. 12 and 13 are flowcharts showing the operation of the embodiment. 1 ... keyboard, 2 ... control unit, 3 ... program memory, 4 ... input buffer, 5 ... kana-kanji conversion dictionary, 6 ... independent word dictionary, 7 ... part-of-speech table, 8 ... sending table, 9 …… Kanji / hiragana notation table, 10 …… Display device.

Claims (1)

(57) [Claims] 1. A dictionary means for storing a large number of words including kanji together with their reading information, and adding and discriminating a word having a fluctuation in notation with said identification code; A kana character storage means storing a plurality of bit patterns representing the presence or absence of a kana character at each character position as a table, and a kanji character storing a plurality of types of kanji characters at each character position / bit patterns representing hiragana notation as a table. Plain display storage means, reading means for reading a word corresponding to the read information input from the input means from the dictionary means, and when the identification code is attached to the word read by the reading means, The notation of the word is sequentially converted to the plurality of types of the notation of the word based on the table stored in the word storage device, and the kanji is changed according to the identification code. Conversion means for converting a part or all of a plurality of kanjis of the word into a hiragana notation based on a table stored in the hiragana notation storage means; A kana-kanji conversion device comprising display means for displaying.