KR100205956B1 - Language code translation device and method - Google Patents

Abstract

1. TECHNICAL FIELD OF THE INVENTION

Interconverter and method for mutually different language codes using hashing function.

2. The technical problem to be solved by the invention

By repeatedly storing the Hangul-Hanja dictionary and the Hanja-Hangul dictionary, it is intended to solve the conventional problems of slowing down the conversion speed and the need for a lot of storage space.

3. Summary of Solution to Invention

Computes one hashing value by combining the corresponding syllable pairs of different language codes, and stores each word to be converted using the same. When a word of the first language code to be converted is input from the user, it is converted into syllable units. After generating the candidate second language code corresponding to each syllable, the hashing value is calculated for each syllable, and the hashing value corresponding to the first syllable among the hashing values is used as an index key. It is configured to search for a matching hashing value, and then find a portion of the matching portion that matches the next hashing value and restore the original language code.

4. Important uses of the invention

Used for Hangul-Hanja interchange converter such as word processor and personal terminal.

Description

Interconversion device and method of different language code using hashing function

The present invention relates to a mutual translator and a method of mutually different language codes (eg, Hangul, Hanja) using a hashing function. In particular, a hashing of language codes corresponding to each other in order to convert between Hangul and Hanja is performed by one syllable unit. After taking a function, by using a hash value calculated by mixing each corresponding hash value of the different languages, the memory space is greatly reduced by storing in a trie structure for each word. The present invention relates to a mutual translator of different language codes and a method thereof.

The Hangul-Hanja interchange converter used in the conventional portable terminal, text editor, desktop publishing system, electronic publishing system, etc. uses the Hangul-Hanja dictionary and the Hanja-Hangul dictionary in order to convert Hangul and Hanja. There is no special compression technique for storing Chinese characters.

For this reason, there is a problem in that it requires a lot of storage space to store the Hangul-Hanja dictionary and the history-Hanja-Hangul dictionary. For example, to store defeats and panes, you need at least 32 bytes of storage.

In addition, the number of string comparisons between Hangul and Hanja words is high, and both the dictionaries need to be searched for, so it is not only time-consuming to convert, but also the dictionary management algorithm has to be complicated when adding or deleting words. .

Accordingly, the present invention devised to solve the problems of the prior art as described above takes a hashing function of the language code corresponding to each syllable unit for mutual conversion between Hangul and Hanja, and then corresponds to the Hangul and Hanja syllables. By mixing the hashing values of each and calculating the final hashing value for each syllable, and storing them in memory in a tri-structure for each word by using them, the space of the memory can be greatly reduced, and the conversion time can be shortened and It is an object of the present invention to provide an apparatus and a method for converting mutually different character codes, which are easier to manage.

1 is a view illustrating a process of generating a hashing value by mixing different Hangul and Hanja one syllable character codes according to the present invention;

2 is a view for explaining a process of restoring Hangul from a hashing value according to the present invention;

3 is a view for explaining a process of restoring Chinese characters from a hashing value according to the present invention;

4 is a schematic block diagram of a Hangul-Hanja interconversion device according to the present invention;

5 is a flowchart illustrating a Hangul-Hanja interconversion method according to the present invention;

* Explanation of symbols for main parts of the drawings

11: Hangul-Hanja Mapping Table

12: Hanja-Hangul conversion unit

13: Hangul, Hanja hash value calculation unit

14: Hangul-Hanja hashing value storage unit

15: Hangul, Hanja Restoration Department

In order to achieve the above object, a mutual translator of different language codes according to the present invention includes a word formed by using hashing values calculated by combining syllables of a second language code corresponding to a syllable having a first language code. First storage means for storing, second storage means for mapping and storing syllables of the second language code corresponding to the syllable of the first language code, and a syllable of the second language code input from the user. A conversion means for converting the corresponding first language code into syllables, and a word of the first language code input from a user in syllable units, and each of the first language code words input from the second storage means. Receiving syllables of a candidate second language corresponding to syllables, or receiving words of a second language code input by a user in syllable units, and converting them into the converting means. A hashing value calculation means for receiving a candidate first language code corresponding to a syllable unit corresponding to the word of the second language code and calculating a hashing value of the syllable unit for the input word; and a language code word input from the user. Restoring means for inputting syllable units, receiving hashing values extracted from the first storage means using hashing values generated by the hashing value calculating means as index keys, and restoring the hashing values to original language codes and outputting the hashing values; And a control means for separating words of a language code input from the user by syllable units, and performing access control and memory management of the first and second storage means.

In addition, the mutual conversion method of different language codes according to the present invention, in the method of interconversion of different language codes applied to the device for mutual conversion of different language codes, combining the corresponding syllable pairs of different language codes After calculating a hashing value, the first step of storing each word to be converted using the same, and if the word of the first language code to be converted from the user is input, it is divided into syllable units to correspond to each syllable. A second step of generating a candidate second language code, a third step of calculating a hashing value for each syllable by using the first language code separated by the syllable unit and the candidate second language code corresponding thereto; And a fourth step of searching for a matching hashing value among the stored hashing values by using a hashing value corresponding to a first syllable among the hashing values as an index key. A fifth step of searching for a record having a hashing value that matches the hashing value corresponding to the second syllable among the records of the search result if the hashing value matching the, and the search result for each syllable of the input word And extracting a portion having the same hashing value, and restoring and outputting the original language code.

Hereinafter, the present invention described with reference to the accompanying drawings will be described by way of example only for the mutual conversion of Hangul and Hanja, but the scope of the present invention is not limited thereto.

1 is a diagram illustrating a process of generating a hashing value by mixing a character code of a Hangul and Hanja one syllable according to the present invention.

When a hashing function is used to mutually convert words of two languages having different code areas, the hashing function proposed in the present invention may be expressed as Equation 1.

[Equation 1]

는 코드 영역이 다른 A라는 언어와 B라는 언어를 서로 혼합하여 C라는 코드 영역으로 변환하는 해싱함수를 나타내며, 이는 충돌(collision)이 없는 완전 해싱함수이다. 또한,

는 해싱값 C로부터 원시 언어 코드인 A나 B를 복원하는 복원 함수를 나타낸다. 이 함수는 해싱함수

의 역함수로서,

이 완전 해싱함수이므로 반드시 1대1로 대응하는 출력이 존재한다.here,

Denotes a hashing function that converts a language A and another language B into a code region C, which is a complete hashing function without collision. Also,

Denotes a reconstruction function that restores the native language code A or B from the hash value C. This function is a hashing function

As an inverse of,

Since this is a complete hashing function, there is a one-to-one corresponding output.

Thus, by using a hashing function and a reconstruction function that satisfy the above conditions, it is possible not only to implement a translator between two languages within the same code system such as KSC-5601 or Uni-code, but also between languages of different code systems. Interconverters can be implemented.

Depending on the code layout of the language you want to convert, you can create various hashing and recovery functions. Since the process of mixing A code domain and B code domain and mapping to C code domain is the same as the already proven pig-hole problem, a hashing function and a reconstruction function exist. [Reference, Udi Manber, Introduction] to Algorithems-A Creative Approach, Addison Wesley, 1989]

The following conditions are assumed in the conversion process of Hangul and Hanja.

First, when converting Hangul to Hanja, one side of the code is already known. That is, in the case of Hangul-Hanja conversion, the corresponding Hanja is found while the Hangul code is known, and in the case of Hanja-Hangul conversion, the opposite is the case.

Second, the words in the Hangul-Hanja conversion dictionary are based on the KSC-5601 code. That is, since only 2350 Hangul characters and 4888 Hanja characters are used, the code area of Hangul and Hanja is quite limited.

Third, in the case of homonyms, the Hangul is the same but the corresponding Hanja codes are different. Therefore, mixing and hashing them does not matter because different hashing values are output.

The code layout of 2 bytes of Korean, Chinese and special characters specified in the KSC-5601 code standard is shown in [Table 1].

TABLE 1

KSC-5601 Korean, Chinese and Special Character Code Areas

KSC-5601 High byte Lower byte Character count Special Characters 0xA1-0xAC 0xA1-0xFE 1128 characters Hangul 0xB0-0xC8 0xA1-0xFE 2350 characters Chinese character 0xCA-0xFD 0xA1-0xFE 4888 characters

Referring to the drawing, the process of making the Chinese character 'Sa (0xDEE4)' corresponding to Hangul 'Sa (0xBBF7)' into one hashing value will be described first. Subtract (0xA1A1) to calculate the Hangul hashing value and the Hanja hashing value. Then, the hashing value '0x5799' is calculated by adding the Hangul hashing value and the Hanja hashing value.

This hashing value is stored in the electronic dictionary for translating Hangul and Hanja. That is, the hashing values of the codes of the one syllable corresponding to each of the Hangul and the Hanja as described above are stored for each word in a trie structure, which is one of data storage structures storing data in units of phonemes. For example, when storing Hangul 'School' and Hanja '學校' as hashing values, Hashing values are calculated for Hangul 'Hak' and Hanja '學' and Hangul 'Gyo' and Hanja '校' respectively. When the word 'school' is stored, the hash value of the 'school' is connected to the hash value of the 'school'.

During the Hangul-Hanja and Hanja-Hangul conversion, the electronic dictionary is searched by the calculated value through this hashing function. The search results consist of hashing values that contain Hangul and Hanja. To restore the original string from this, use the inverse function of the hashing function.

2 is a diagram illustrating a process of restoring Hangul from a hashing value.

To recover Hangul from a hashing value, you must know the hashing value, hashing constant (0xA1A1), and the Hanja you want to convert.

In other words, when restoring Hangul from a hashing value, the hashing value is added to the hashing value twice, and then the code value of the Hanja syllable that is inputted from this value is subtracted.

FIG. 3 is a diagram illustrating a process of restoring Chinese characters from a hashing value. Similar to FIG. 2, when restoring Chinese characters from a hashing value, a hashing constant is added to a hashing value twice, and then a Hangul one syllable is input. You can subtract the code value.

If this is expressed as a formula of a hashing function and a restoring function, Equation 2 is expressed.

[Equation 2]

는 해싱 값,

은 한글 코드 값,

는 한자 코드값, 0xA1A1은 해싱 상수를 나타낸다.here,

Is the hashing value,

Is the Hangul code value,

Is a Chinese character code value, and 0xA1A1 is a hashing constant.

4 is a schematic block diagram of a Hangul-Hanja interconversion apparatus according to the present invention, where '11' is a Hangul-Hanja mapping table, '12' is a Hanja-Hangul conversion unit, and '13' is a The Hangul and Hanja hashing value calculator, '14' represents the Hangul-Hanja hashing value storage unit, and '15' represents the Hangul and Hanja restoration units, respectively.

The present invention is applied when converting between Hangul and Hanja for a word instead of one syllable, and performing conversion as between conventional Hangul and Hanja for one syllable.

In order to convert between Hangul and Hanja, first, a hashing value of the Hangul and Hanja syllables corresponding to the Hangul-Hanja hashing value storage unit 14 is calculated and these hashing values should be stored in a tri-structure for each word.

First, the Hangul word is input, the input Hangul word is divided into one syllable unit, and candidate Chinese characters for each syllable are obtained through the Hangul-Hanja mapping table 11 as in a general method supported by all word processors. The candidate Hanja for each syllable and the syllable of the input Hangul word are calculated by hashing the Hangul and Hanja syllables inputted by the Hangul Hanja syllable calculator 13 and calculating the hashing values for each Hanja. Create a value.

Among the hashing values calculated by the Hangul and Hanja hashing value calculator 13, first, a plurality of hashing values for the first syllable of the input Hangul word are searched in the Hangul-Hanja hashing value storage unit 14, Then, the searched hashing value is searched for a hashing value corresponding to the second syllable among the hashing values connected by the tri structure.

When the corresponding hashing values are extracted through this process, the Hangul and Hanja restoring unit 15 outputs the original Chinese word to the user using a restoration function.

In the case of a Hanja word, candidate Hanguls for each syllable of the Hanja word input by the Hanja-Hangul conversion unit 12 are generated and input to the Hangul and Hanja hash value calculation unit 13.

The Hangul and Hanja hashing value calculator 13 calculates hashing values for the syllables and candidate Hangul syllables of the input Hanja word and converts the Hangul word through the same process as the Hangul-Hanja conversion. Print to user

The above operation process will be described using the word 'school' as an example.

If the user wants to convert Hangul to 'Hanja' by inputting Hangul 'School', when the user enters Hangul 'School', the Hangul-Hanja interchange conversion control unit converts the word 'school' into syllable units of 'Hak' and 'Kyo' Separate.

Then, using the Hangul-Hanja mapping table 11, five candidate Chinese characters for 'hak' are obtained, hashing is performed on five candidate Chinese characters and Hangul 'hak', and five hashing values are calculated. Similarly, using the Hangul-Hanja mapping table 11, 25 candidate Hanja characters for Kyo are obtained to calculate 25 hashing values.

In the Hangul-Hanja hash value storage unit 14, the hashing value for 'hak' and the hash value for 'kyo' are connected in a tri-structure, so there is only one connection for the word 'school'. That is, when the first syllable is 'hak', there may be five hashing values. However, when the first syllable has a hashing value for 'school' connected to 'hak', there is only one case that satisfies the meaning of 'school'. If there are a plurality of Hanja words for the input Hangul word, a plurality is output to the user.

Here, the inverted Hangul-Hanja mapping table and hashing dictionary are stored in general memory. This part can be constructed using various data structures and search algorithms, such as search using syllable tries, or binary search.

In addition, the Hangul, Hanja hash value calculation unit, and the Hangul, Hanja recovery unit can be implemented using an addition circuit and a subtraction circuit.

In addition, when the Hangul-Hanja mapping table and the hashing dictionary are stored using the associated memory, the Hangul-Hanja conversion speed is faster than other techniques, and the apparatus is simplified because the table search circuit is unnecessary.

5 is a flowchart illustrating a Hangul-Hanja interconversion method according to the present invention.

First, a hashing value is calculated for each pair of Hangul and Hanja corresponding to each syllable unit (101), and each word is stored as a hashing value in a tri-structure using the calculated hashing value (102).

When the Hangul to be converted is input from the user (103, 104), the input Hangul word is divided into syllable units (105), and the candidate Hanja for each syllable is obtained by referring to the Hangul-Hanja mapping table (106).

Then, a hashing value is calculated by combining each Hangul syllable and the candidate Hanja syllables corresponding thereto (109), and the Hangul-Hanja Hashing Dictionary is obtained by using the hashing value corresponding to the first syllable among the obtained hashing values as the first index value. Search 110.

As a result of the search, if a corresponding value exists, it is searched whether there is a value equal to a hashing value for the second syllable among the connected hashing values after the found hashing value (111).

In this way, a search is performed on each syllable of all words inputted in this manner (112). If there is any same value from the Hangul-Hanja hashing dictionary, this value is extracted, and the hashing value is restored to the original Hanja (113). In step 114, the restored Chinese characters are output to the screen.

Here, the searching process is similar to the conventional one, and if there is no Hanja word corresponding to all Korean words input as before, only the matching Hanja word is output to the user.

When the Chinese characters to be converted are input from the user (103, 104), the input Chinese characters words are divided into syllable units (107), and candidate Hanguls for each syllable are obtained through the Hanja-Hangul conversion (108).

In operation 114, the Hangul for the Hanja word input through the same process (109 to 113) as the Hangul-Hanja conversion process is output.

The present invention described above is capable of various substitutions, modifications, and changes within the scope without departing from the technical spirit of the present invention for those skilled in the art to which the present invention pertains, and thus is limited to the above-described embodiments and drawings. It is not.

The present invention made as described above has the following unique effects.

First, the Hangul-Hanja conversion dictionary may be configured as a dictionary, and the storage space may be reduced by 1/4 or more as compared with the conventional art. In other words, in the case of defeat and pain, for example, 32 bytes of storage space are conventionally required, but the present invention requires 6 bytes to represent a tri structure, and 2 bytes are additionally used to maintain a tri node. A total of 8 bytes of storage is needed.

Second, the Hangul-Hanja conversion can be performed much faster than before.

Third, it is easy to hardwareize the Hangul-Hanja interconverter, and when it is added to a personal terminal, it is possible to improve the performance of the Hangul-Hanja translator by storing a larger number of Hangul and Hanja word pairs in a limited size memory. Can be.

Fourth, when the associative memory is used, the Hangul-Hanja conversion speed can be further increased.

Claims (9)

An apparatus for translating different language codes, First storage means for storing a word formed by using hashing values calculated by combining syllables of a second language code corresponding to a syllable having a first language code; Second storage means for mapping and storing syllables of the second language code corresponding to syllables of the first language code; Conversion means for converting a syllable of the second language code input from a user into a syllable of the first language code corresponding thereto; Receives a word of the first language code input from the user in syllable units, receives syllables of a candidate second language corresponding to each syllable of the input first language code word from the second storage means, or the user A syllable unit for a word input by receiving a word of a second language code input from a syllable unit and receiving a candidate first language code of a syllable unit corresponding to a word of the input second language code from the converting means Hashing value calculating means for calculating a hashing value of; Receives a language code word input from a user in syllable units, receives hashing values extracted from the first storage means using hashing values generated by the hashing value calculating means as index keys, and converts the hashing value to the original language code. Restoring means for restoring the output; And And a control means for separating words of a language code input from the user by syllable units and performing access control and memory management of the first and second storage means. The method of claim 1, And said first and second storage means comprise an associative memory. The method according to claim 1 or 2, The first storage means, Inter-converter of different language codes, characterized in that each word is stored by storing a hashing value in a syllable unit tri structure. The method of claim 3, wherein The hashing value calculation means, A hashing constant is subtracted from each of the one syllable code value of the first language code and the one syllable code value of the second language code, and then the summed values of the subtracted first and second language codes are added to hashing. Interconverter of different language code, characterized in that the value is calculated. The method of claim 3, wherein The restoration means, And adding a hashing constant twice to a hashing value output from the first storage means, and subtracting a syllable for a word input from the user to restore an original language code. In the method of interconversion of different language codes applied to a device for translating different language codes, Calculating a hashing value by combining corresponding syllable pairs of different language codes, and storing each word to be converted using the same; A second step of generating a candidate second language code corresponding to each syllable by dividing the word of the first language code to be converted from the user into syllable units; Calculating a hashing value for each syllable by using the first language code separated by the syllable unit and the candidate second language code corresponding thereto; A fourth step of searching for a matching hashing value among the stored hashing values by using a hashing value corresponding to a first syllable among the hashing values as an index key; A fifth step of searching for a record having a hashing value corresponding to a hashing value corresponding to a second syllable among records of the search result if a hashing value matching the search result exists; And And extracting a portion having the same hashing value as a result of searching for each syllable of the input word, and then restoring and outputting the original language code. The method of claim 6, The hashing value is, Subtract a hashing constant from each of the one syllable code value of the first language code and the one syllable code value of the second language code, and then calculate the sum of the subtracted first and second language codes. Interconversion method of different language codes, characterized in that the. The method of claim 7, wherein And a hashing value is stored in a tri-structure of syllable units when the word is stored in the first step. The method of claim 8, The hashing value restoration is, And adding a hashing constant twice to the extracted hashing value, subtracting a syllable for a word input from the user, and restoring an original language code.

Images