JPH0660063A - Chinese character converting device - Google Patents

Abstract

PURPOSE:To provide the KANJI(Chinese character) converting device which eliminates the need to remember the index code of a word that a user has registered and efficiently performs the conversion of the registered word, the registration of a word, and the deletion of the registered word from a file. CONSTITUTION:The KANJI converting device which converts reading KANA (Japanese syllabary) into KANJI performs arithmetic based upon a specific function using an index code by a dictionary rearrangement part 16 to extract a block corresponding to the index code of a desired word in a learning dictionary 17 at a high speed when the registration of a word or the deletion of a registered word from the learning dictionary 17 where the word that a user desires can be registered is performed. A retrieval part 14 also retrieves a word in the block by said corresponding block taking-out method and performs KANA-KANJI conversion of the reading KANA registered in the learning dictionary 14.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a kanji conversion device,
For details, refer to the word registration.

[0002]

2. Description of the Related Art Generally, a basic dictionary of a kanji conversion device stores more than 32,000 everyday words. Of the 32,000 words stored, two-letter words were the most common, with 25,000 words accounting for about 79%, followed by more than 3,000 three-letter words. It is about 10%, followed by about three thousand four-letter words, accounting for a little over 9%, and five to seven letter words accounting for about 0.5%.

However, these more than 32,000 words are all everyday terms, such as words such as "Zhangjiahua" and "Li Daen", and new words such as those shown in 1 and 2 below.

[0004]

[Outer 1]

[0005]

[Outside 2]

[0006] Technical terms such as "stoppage", "financing" and specific terms such as "Matsushita Electric Industrial Co., Ltd." are not registered in the dictionary. As a function of allowing the user to freely register the unregistered word as described above, there is, for example, a dictionary registration function of the Fengtian Chinese sentence system, which has the highest share of the computer market in the Republic of China in the conventional kanji conversion device.

FIG. 9 shows a block diagram of a dictionary registration function unit of a conventional Kanji conversion apparatus. 33 is a dictionary section,
It stores a predetermined word and a registered word registered by the user. An index code setting unit 31 receives a character string or a symbol input by a user and sets the character string or symbol as an index code corresponding to a word to be registered designated by a designation unit (not shown). To do.

Reference numeral 32 denotes a word registration unit which associates a word to be registered designated by a designation unit (not shown) with the index code set by the index code setting unit, and registers them in the dictionary unit 33. Is. An input unit 34 receives the index code input by the user.

Reference numeral 35 is a search unit, which cuts out the index code input from the input unit and outputs the index code to the order search unit 36. An order search unit 36 sequentially searches the dictionary unit 33 based on the index code and outputs the corresponding registered word to the output unit. An output unit 37 outputs a conversion result.

The operation of the word registration function unit of the conventional Kanji conversion apparatus as described above will be described below. Dictionary of words 3
When registering in No. 3, first, the user instructs the word to be registered from a not-shown designation unit. next,
The index code setting unit 31 receives a character string or a symbol input by the user corresponding to the designated word and sets it as an index code. As shown in FIG. 10, the word registration unit 32 associates the specified registration word with the index code, and stores a certain length in the left column in a specific file in the dictionary 33 in which the word is registered. Sano index code, and the registered words corresponding to the index code are stored in the right column in a sequence method. This completes the word registration.

When converting a word registered as described above, when the user inputs an index code corresponding to the word to be converted from the input unit 34, the index code is read and searched. It is output to the order retrieval unit 36 through the unit 35. Based on the index code, the order search unit 36 sequentially searches the registered words stored in a specific file for registering words in the dictionary 33, detects the registered word corresponding to the index code, and outputs this word. It is output to the unit 37.

In addition, when deleting a registered word from the dictionary 33, when the user gives a deletion instruction and inputs an index code corresponding to the word to be deleted,
The index code is output to the order search unit 36 through the search unit 35. The order search unit 36 sequentially searches for a word corresponding to the index code based on the index code, and deletes this word from the dictionary.

The stored contents of the specific file for registering the words are read at the time when the Chinese system is activated.

[0014]

However, the word registration function of the Kanji conversion device as described above has the following problems. (1) In the conventional case, since the index code is determined and used by the user, even if the same word is registered, the index code varies depending on the user, and therefore online registration cannot be performed. Not very flexible in use.

(2) When using a registered word, the user has to remember the index code of the registered word, which is inconvenient for the user and is inconvenient. (3) When the registered words are converted, the search is performed by a sequence method in which the words registered in the dictionary are sequentially searched and the word corresponding to the index code is detected. , The search response time will increase,
Operability is reduced.

(4) Regarding registration of a word and deletion of the registered word from the dictionary, storing in a file in which the word is registered and searching for a file in which the word is registered are performed in a sequence method. As it is done, it takes time to register and delete words, and the more registered words, the more inefficient it becomes. In view of the above problems, the present invention does not require the user to remember the index code of the registered word, and furthermore, the conversion of the registered word, the registration of the word, and the file of the registered word. An object of the present invention is to provide a kanji conversion device that can efficiently delete characters from.

[0017]

In order to solve the above-mentioned problems, in the present invention, a learning dictionary in which a word desired by a user is registered in a kanji conversion device for converting a word index code into kanji. And a dictionary organizing means for registering and deleting words in the learning dictionary based on a value calculated by a specific function using the index code, and a specific function using the index code. A conversion unit that searches a word registered in the learning dictionary based on the value obtained and converts the word into a Kanji corresponding to the index code.

[0018]

In the kanji conversion device for converting the word index code into kanji according to the present invention, the user's desired word is registered in the learning dictionary. The dictionary organizing means registers and deletes words in the learning dictionary based on the value calculated by the specific function using the index code. The conversion means, based on the value calculated by the specific function using the index code,
The words registered in the learning dictionary are searched and the index code is converted into kanji.

[0019]

DESCRIPTION OF THE PREFERRED EMBODIMENTS FIG. 1 is a block diagram of an embodiment of a kanji conversion device of the present invention. In FIG. 1, reference numeral 11 denotes an input unit that receives a vocalization symbol (search symbol) of an arbitrary length input by the user.

Reference numeral 12 denotes an index code processing unit, which performs a process of converting a voicing symbol input by a user into a basic dictionary, and converting it into an index code used for a hash operation to be described later. Reference numeral 15 is a basic dictionary, which stores predetermined words. A learning dictionary 17 stores words registered by the user.

More specifically, as shown in FIG. 7, the learning dictionary 17 has a plurality of fixed lengths, a related block information column 171, an available space information column 172, and a registered word index code column 173. , And the corresponding character string code column 17
It is composed of four parts, and the following information is stored in each part. That is, the related block information column 171
Stores the address of the related block in which the next word is stored when the available space, which is a space for registering a word, runs out.

The available space information column 172 stores available space information indicating how many bytes of the available space remain. The registered word index code column 173 and the corresponding character string code column 174 store the index code of the registered word and the character string code corresponding to the index code.

Reference numeral 14 is a search unit, which searches the basic dictionary 15 based on the index code, and the learning dictionary 17
With respect to, the search is performed based on the address of the block derived by the hash function described later based on the index code. Reference numeral 13 is a conversion unit, which cuts out the conversion range index code according to the specification of the conversion range by the user.
Based on the search result of the search unit 14, the character string corresponding to the index code of the conversion range is converted.

Reference numeral 20 denotes a storage unit, which is used for calculation of an address of a block in which a word is stored when converting into Kanji, registering a word, and deleting a registered word from a learning dictionary. It stores data and data used for retrieval. A conversion correction unit 19 corrects the conversion result by designating a conversion correction key.

Reference numeral 16 is a dictionary organizing unit, which is a learning dictionary 1 derived by a hash function described later based on the index code.
Based on the addresses of the blocks in 7, the words are registered, the registered words are deleted from the learning dictionary, and the dictionary is reorganized. An output unit 18 outputs a conversion result and a correction result. Here, before explaining the operation of the Kanji conversion apparatus, the operation of the index code processing unit 12 will be described with reference to FIGS.

Chinese readings are indicated by a combination of phonetics, votives, votives, and tones, and there are over 1,300 valid readings. When encoding the phonetic symbols that represent these readings, two bytes are required to represent each reading. For example, the Chinese vowel and the Chinese vowel of each Chinese phonetic symbol are placed together in one byte, the vowel and tone are placed in another byte, and are placed in the ASCII character table as shown in FIG. . Then, all the readings can be represented by a 2-byte code composed of the combination of FIGS. 6A and 6B, and this code is the index code.

The process of converting the input utterance symbols into the index code as described above uses the order values given to the Chinese utterance symbols shown in FIG. 5, and as shown in FIG. It is done by simple judgment and calculation. First, Byte1 and Byte2, which are the values of the first and second bytes of the index code, are set to zero (S101). Next, it is determined whether or not the input voicing symbol has a vowel (S
102), if there is a vowel, the operation of Byte1 + 21H is performed, and the value obtained here is set as Byte1 (S).
105). Based on the voice vowel code order shown in FIG. 5A, the corresponding value of the voice vowel is extracted (S106), and the voice vowel corresponding value is multiplied by 4 to obtain the By obtained in S35.
In addition to te1, this time, the value obtained here is Byte1
(S107). Next, based on the intermediary code order shown in FIG. 5B, the corresponding value of the intermediary is extracted (S104), Byte1 obtained in S37 is added, and the final value of the first byte. Was obtained (S108).

In S102, if the input voicing symbol has no vowel, first, Byte1 + 75H
Is calculated, and this value is set to Byte1 (S10
3) Next, the processes of S104 and S108 are performed to obtain the final value of the first byte. After the value of the first byte is obtained, the process of obtaining the value of the second byte is subsequently performed.

First, it is judged whether or not the input vocalization symbol has a vowel (S109).
+ 26H is calculated and the value obtained here is Byte2
(S112). Based on the order of the verbs shown in FIG. 5C, the corresponding values of the verbs are extracted (S
113), the value corresponding to the phonetic symbol is multiplied by 5, and added to Byte2 obtained in S112, and the value obtained here is set as Byte2 (S114). Next, based on the tone code order shown in FIG. 5D, the corresponding value of the tone is extracted (S111), Byte2 obtained in S114 is added, and the final value of the second byte can be obtained. It was possible (S115).

Finally, the codes corresponding to the first and second byte values are searched from the ASCII character table shown in FIG. 6, and the searched codes are combined and output as an index code (S116). In S109, if the input vocalization symbol does not have a vowel, first, B
yte2 + 21H is calculated, this value is set to Byte2 (S110), then the processes of S111 and S115 are performed, and the index code is output (S116).

In the above description and the flowchart of FIG. 4, H is added after the hexadecimal number in order to distinguish the decimal number from the hexadecimal number.
The operation of the Kanji conversion device having the above configuration will be described below with reference to the drawings. First, the kanji conversion operation will be described.

When the user inputs an utterance symbol from the input unit 11 (S1, S2), the input utterance symbol is stored in the storage unit 20. Next, it is determined whether the vocalization symbol is complete (S23). The phonetic symbols of the Chinese characters include "Ryomi, Tone", "Koma, Ryomo, Tone",
"Voice Mother, Rhyme Mother, Tone", "Voice Mother, Caregiver, Rhyme Mother, Tone",
There are five types of "voice mother, tone". If the input phonetic symbol has one of these five types, it is determined that the phonetic symbol is complete. When the tone is input, the reading of one character is finished.

If it is determined in S23 that it is complete, the index code processing unit 12 converts the vocalization symbol stored in the storage unit 10 into the corresponding index code by the procedure described above (S24). The converted index code is stored in the storage unit 20. Then, when the range of reading to be converted is designated by the user, the index code corresponding to the conversion range is cut out from the index code (S25),
The cut out index code is passed to the search unit 14 to perform the search operation.

First, the search unit 14 searches the basic dictionary 15 stepwise and by a binary search method using the passed index code as a search key (S26). Then, the learning dictionary 17 is searched as follows. That is, first, the storage unit 2
The index code stored in 0 is read (S27), and the calculated hash code is used to obtain the block position in the learning dictionary corresponding to the reading using the read index code (S28). ).

The method of calculating the hash function of this embodiment is based on the ASCII values of the three codes before the index code string. For example, the index code string (a ₁ , a ₂ ,
a ₃ , a ₄ ,. ． ． ． ． ． The position H ₁ (X) of the block of the learning dictionary 17 corresponding to () can be obtained by the following formula. H ₁ (X) = (a ₁ * b ³ + a ₂ * b ² + a ₃ * b ¹ ) modN where b is a specific integer and N is a preset basic number of blocks. mod is an operator for calculating the remainder of division.

The value calculated by the above-mentioned hash function is taken as the address of the corresponding block in the learning dictionary 17, and the block is taken out (S29). The contents of the extracted block are stored in the storage unit 20 (S30). Then, it is checked whether or not there is data in the index code column 173 and the corresponding character string column 174 of this block (S3).
1). If there is no data, it is determined that the block is an empty block, and the next related block is read according to the address of the related block stored in the related block information column 171 (S34).

If it is not an empty block, all the index codes stored in this block are compared with the order in S33 (S32), and if there is a matching index code (S32).
33), the corresponding character string is passed to the conversion unit 13 and displayed on the output unit 18 in S36. If the relevant word is not found (S33), the next related block is read out by the address of the related block in the related block information column 171 (S3).
4). The address of the related block read in S34 is checked, and if it is zero, it is determined that this block is the last block, and a message indicating that there is no corresponding word is output (S36).

If it is not the last related block, the processing from S29 to S35 is repeated with the message that the corresponding word is detected or the corresponding word is not found output. Further, when the conversion result is corrected, first, when the user inputs a conversion correction instruction (S1 to S1).
3), the conversion correction unit 19 outputs another change candidate (S).
4). When the user selects a conversion candidate from the output conversion results, the conversion correction unit 19 corrects the conversion result and outputs it.

Next, the operation of reorganizing the dictionary such as registration and deletion with respect to the learning dictionary 17 will be described. First, when registering a word, when the user designates the word to be registered (S1 to S3), the dictionary organizing unit 16 reads the index code of the word (S7), and the hash function is applied to the above index code. Is calculated (S8) to obtain the address of the corresponding block in which the word is registered.

The calculation method of the hash function is the same as above. The corresponding block is read from the address based on the operation result of the hash function (S9), and it is checked whether the space for registering this word in the block, that is, the available space is sufficient (S11). If the available space is sufficient, it is next determined whether the independent space is sufficient (S1).
3).

For example, to store 8 bytes, FIG.
When the independent space is sufficient as in (a), the word is registered in the first open space (S15). on the other hand,
Although the available space is sufficient, if the available space is formed from disjoint independent spaces and a new word cannot be registered as it is, for example, when trying to register an 8-byte word, FIG. As shown in, the 10-byte space still remains in the block, but the remaining 10-byte space is not continuous, and 6 and 4 respectively.
If it exists in the form of two independent spaces of bytes, registration is performed as follows.

That is, the dictionary organizing unit 16 moves the already registered word to a lower address, connects the independent spaces to form a 10-byte space (S14), and then puts the word in this space. Register (S15). In addition, when there is no available space in the acquired block, as shown in FIG.
As shown in (c), the dictionary organizing unit 16 uses the following H
_2. Form a new block with associated address value calculated by the (X) function, i.e. the associated block.

H ₂ (X) = M increment 1 Here, M is a variable, which is the highest block position immediately before, which is the current total number of blocks. The initial value of M is N-1. The address value of this related block is stored in the related block information column 171. Next, in S15, a new word is registered in this related block, and the word registration is completed.

When deleting a word registered in the learning dictionary 17, when the user specifies a word to be deleted, the dictionary organizing unit 16 reads out the index code in the specified range (S7). By the same hash function operation as described above, the address value of the corresponding block corresponding to the index code is obtained, and the content of the corresponding block is read (S1
0). It is searched whether or not there is a word to be deleted in the read corresponding block (S16). When the word is found (S17), the storage space is released, and the registered word is deleted from the learning dictionary (S18).
If the word does not exist (S17), it is determined whether or not there is a related block (S19). If not, the word is output to the output unit that it is not registered (S2).
0). If there is a related block, S16 to S2 are executed until the word is found or the related block disappears.
Repeat operation 1).

In order to further understand the operation of the above-mentioned kanji conversion device, Chinese voicing symbols shown in the following 3 to 11 are shown.

[0046]

[Outside 3]

[0047]

[Outside 4]

[0048]

[Outside 5]

[0049]

[Outside 6]

[0050]

[Outside 7]

[0051]

[Outside 8]

[0052]

[Outside 9]

[0053]

[Outside 10]

[0054]

[Outside 11]

A specific operation of the kanji conversion device will be described below by using the example of converting "to" into "coral troops of marine corridor". When the user inputs the above vocalization symbols (outer 3) to (outer 11) from the input unit 11 (S1,
2) The phonetic symbol is stored in the storage unit 20 (S2).
2). Next, the index code processing unit 12 performs index code conversion processing based on FIG. 4 on the stored utterance symbols to convert the utterance symbols into the following index codes,
It is stored in the storage unit 20.

As described above, the index code processing unit 12
Converts the vocalization symbol into an index code as follows with reference to the sequence value in FIG. That is, the voice mother shown in the outer 12

[0057]

[Outside 12]

[0000] is the 10th in the vowel order, and there is no messenger, so it is the 0th in the vowel order, so the 1st byte of the index code is as follows. 21H + 10 * 4 + 0 = 49H, 49H corresponds to "I" of the ASCII code. In addition, the following final 13

[0059]

[Outside 13]

Is the fourth item in the prosodic order, and
Tone shown in

[0061]

[Outside 14]

Since is the second in the tone order, the second byte of the index code is as follows. 26H + 4 * 5 + 2 = 3 cH, 3 cH corresponds to the ASCII code “<”. From the above result, the index code of the vocalization symbol (outer 3) is "I <".

By performing the same processing, the above-mentioned vocalization symbols (outer 3) to (outer 11) become the index code "I <aN%.
= "] NVC] E0w? 53" (S2
4). When the conversion range is designated by the user, the conversion unit 13 cuts out the conversion range (S25), and the search unit 14 searches the basic dictionary based on the index code (S26).
Then, the learning dictionary is searched based on the block position calculated by the hash function as described above.

Here, "(outside 5) (outside 6) troops", "(outside 7) (outside 8) advance", "(outside 9) (outside 1" are added to the basic dictionary.
It is assumed that the three words "0) (External 11) Kaeui Special" are registered, and the personal name word "(External 3) (External 4) Kaicor" is not registered in the basic dictionary or the learning dictionary. . In this case, the search unit 14 detects the above-mentioned three registered words from the basic dictionary 15 (S26), but the corresponding words (outer 3) and (outer 4) are not registered, so that
With respect to the readings (outer 3) and (outer 4) cut out in S25, the first candidate characters “Umi” and “Mountain” of the homophones are output.

Then, the conversion result becomes "Kaiyama Squadron Advancement Special", and the display unit 18 outputs the conversion result. However, in this case, since the conversion of "Marine troop soldiers is special" is performed, when the user inputs a conversion correction instruction to correct the erroneous conversion of "mountain", the vocalization symbol (outer 4) As synonyms, mountain, 衫, fan, cedar, coral,
Kanji, and the following Kanji characters shown in Outer 15 and Outer 16 are output.

[0066]

[Outside 15]

[0067]

[Outside 16]

When the user selects the desired character "coral" from these homonyms, the conversion result is corrected, and the correct conversion result "sea coral troops attacking speciality" is output by the output unit 18. Is output. At this time, if the word "sea coral" is registered in the learning dictionary 17 by the dictionary organizing unit 16, erroneous conversion in the future can be avoided.

First, in the case of registration, the vocalization symbol of the word to be registered is converted into a corresponding correct character string, and then registration is started. When the range of words to be registered is designated by the user, the dictionary organizing unit 16 reads the index code of the designated range (S7) and stores the index code in the storage unit 20. Next, the hash function is operated to extract the position of the corresponding block of the index code stored in the storage unit 20 (S8).

A specific calculation is performed by using three codes (I is 7) before the index code “I <aN” stored in the storage unit 20.
3, <60 is used, a is 97), and N is 64, b
17 and the storage space of each block is 1024 Kbytes, the result is as follows. H ₁ (X) = (73 * 17 ³ + 60 * 17 ² + 97 * 17) m
od64 = 377638 mod64 = 38 Here, the block at the block position 38 extracted based on the calculation result of the hash function is as shown in FIG.
If the available space is only 4 bytes and it is not possible to register an 8-byte word, the dictionary organizing unit 16 obtains the position of the related block by the H ₂ (X) function as described below.

H ₂ (X) = 64 The calculated address value is stored in the related block information column 171. Next, the new word is registered in the related block whose block position is 64 (S15). If "(outside 3) (outside 4) sea coral" is registered in the learning dictionary 17 and the above vocalization symbol is input again, after searching the basic dictionary 15,
As described above, the learning dictionary 17 is searched by the search method using the hash function, and “sea coral” corresponding to (outer 3) and (outer 4) is output.

As a result, the correction result does not need to be corrected, and it is possible to accurately convert the correction result into the "sea coral troops advancing speciality". The present invention is not limited to the above embodiment,
Modifications can be made without changing the gist. For example, the symbol to be input is not limited to the vocal symbol, and the simple Kuramo symbol may be used.

Further, the conversion rule of the index code of the input symbol is not limited, and it is assumed that the phonetic and tone of each Chinese reading are placed together in one byte, and the messenger and vowel are placed in another byte. Also, the conversion operation of the index code processing unit may be modified. In this embodiment, a three-dimensional hash function of a polynomial is calculated by three codes of the index code,
It is also possible to implement all index codes as the calculation target of the hash function. All the above-mentioned methods belong to the present invention.

[0074]

As described above, according to the present invention, by providing the converting means, the dictionary organizing means, etc., the index code corresponding to the word registered by the kanji converting device is determined, so that the user can make an index one by one. You don't have to remember the code and you can register online.

Further, since the position where the registered word is registered is calculated from the index code and the search is performed, the response time of the search is shortened and the conversion efficiency is improved. From the above, there is an effect that operability is greatly improved for conversion of Chinese characters such as Chinese and Japanese.

[0076]

[Brief description of drawings]

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

FIG. 2 is a diagram showing a flowchart of processing steps of the Kanji conversion apparatus of the present invention.

FIG. 3 is a diagram showing a flowchart of processing steps of the kanji conversion apparatus of the present invention.

FIG. 4 is a diagram showing a flowchart of an operation of an index code processing unit.

FIG. 5 is a diagram showing a code sequence of Chinese vocalization symbols.

FIG. 6 is an explanatory diagram showing coding of vocalization symbols in Chinese according to the embodiment of this invention.

FIG. 7 is a diagram for explaining the detection of the position of the corresponding block by the structure of the learning dictionary and the index code according to the embodiment of the present invention.

FIG. 8 is an explanatory diagram illustrating a registration state of the learning dictionary according to the embodiment of this invention.

FIG. 9 is a block diagram showing a configuration of a conventional kanji conversion device.

FIG. 10 is a diagram showing a configuration of a dictionary of a conventional example.

[Explanation of symbols]

 12 index code processing unit 13 conversion unit 14 search unit 16 dictionary organization unit 17 learning dictionary

Front Page Continuation (72) Inventor Kei Ken Hayashi Taiwan Taipai Shi Tha An Chi 10628 Ren Ile San Toan 136 Hao 10 Rosun Sha Tien・ Chi Chi Shu Kai Far Khu Feng Yu Xieng Kong Su (72) Inventor Zhang Jiahua Taiwan Taipai She Tha An Chi 10628 Ren In Ile San Touan 136 Hao 10 Lo Sun Sha Tien Chi Chi Shu Kai Far Khu Feng Yu Si Xen Con Su

Claims (1)

[Claims] 1. A kanji conversion device for converting a word index code into kanji, based on a learning dictionary in which a user's desired word is registered and a value calculated by a specific function using the index code. , A dictionary organizing unit for registering and deleting words in the learning dictionary, and a word registered in the learning dictionary based on a value calculated by a specific function using the index code,
A kanji conversion device comprising: a conversion means for converting a kanji corresponding to an index code.