JPH0760433B2 - Kanji converter - Google Patents

Description

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a kanji conversion device for converting a phonetic character string such as Chinese into a kanji character string.

Conventional technology In Chinese, in principle, one kanji corresponds to one syllable. The syllable is composed of a mother, a mother and a tone as shown below.

Table 1 shows the phonological elements of Chinese in the phonetic system used in the Chinese mainland (Hai-on 1) and Taiwan (Hai-on 2, note), respectively. The phonetic method of No. 1 will be used for explanation). Columns (1) to (21) in the table are called a mother, and columns (22) to (59) are called a mother. Among them, especially column (22) ~
The melody of (24) will be called as a phoneme, the melody of columns (25) to (37) will be called a main melody, and the melody of columns (38) to (59) will be called a combined melody. The combined syllable is composed of a combination of the consonant and the main syllable. In other words, the notation symbols i, u, yu of the consonants and the notation symbols a, o, e, ai, ...
Although it does not necessarily consist of a combination with eng, the analysis of the phonological elements that compose them actually reveals that the phonological elements of the consonant and the main vowel are included. I'm out. For example, “ui” includes phoneme elements “u” and “ei”. Table 2 shows the phoneme elements of the consonant and main vowel included in each combined vowel. Therefore, it can be said that the Chinese syllables are composed in the order of mother tongue + consonant + main rhyme + tone. For example. "Time""Yi"
The phonological elements included in the reading of kanji such as "love,""take,""hatsu,""forget," and "medium" are as follows.

As can be seen from this example, each syllable does not necessarily include all of the mother, the consonant, and the main syllable. For example, the reading "shi '" of "time" includes only the phoneme element of the mother of voice.
Also, the tone is an essential element in the construction of each syllable. By this, the height of the syllable tone can be known.

The phoneme elements separated by dotted lines in Table 1 are those whose pronunciations are most similar to each other and are often mistaken. For example, both “ch” and “c” are affricate sounds, which are distinguished only by the position of the tongue when pronounced, which is very difficult for a person who learns Chinese. In particular, in a kanji conversion device that uses phonetic characters as input means, the user can convert the kanji string corresponding to the kanji string that the user wants to input into the kanji string by inputting the syllable corresponding to the specified phonetic character string. If similar phonological elements are not correctly distinguished and input, erroneous conversion or non-conversion becomes possible, which is a fatal drawback for a kanji conversion device.

An example of a conventional kanji-to-kanji conversion device is Japanese Patent Laid-Open No. 59-
No. 121425 is disclosed. FIG. 2 shows a block diagram of this conventional syllable-kanji conversion device. Reference numeral 21 is a separating means for separating the input data into Roman character data and tone data. Reference numeral 23 is a dictionary that stores the Roman character string, the Chinese character string, the tone and the frequency of use for each word as shown below.

Reference numeral 22 is a reference means for extracting from the dictionary 23 all homonyms corresponding to the Roman character string data given by the separating means 21. Reference numeral 24 compares the kanji string obtained from the reference means 22 with the tone data of the separating means 21 and outputs a predetermined kanji string, and if there is no tone data, the frequency of use of the corresponding kanji string is used frequently. It is a comparison means that outputs in sequence and can select a desired kanji string.

In the conventional syllabary-kanji conversion device configured as described above, for example, when "Chinese" is desired to be input, the reading "zhong1 guo2" is first input from the keyboard. Then, the separation means 21 separates the Roman character string data of (zhongguo) and the tone data of (1,2). Reference means 24
With (zhongguo) as a search key, words are sequentially searched from the dictionary 23. The word registered in (zhongguo) in the dictionary 23 is "China" However, it is "China" that the voice data is (1,2).
Therefore, the comparison means 24 determines that “China” is output.

Problems to be Solved by the Invention However, the above configuration has the following problems.

(1) When the user cannot distinguish similar phoneme elements as shown in Table 1, for example, when he wants to input "student", the correct reading is If you don't have a clue, enter the wrong reading,
Can't be converted correctly as "student". In such cases, there is no choice but to try all the possible combinations one by one, as if by trial and error.

(2) There are about 1260 types of Chinese kanji readings, and if you code them, you can use at most 2bytes (when using bytes as a unit), but the dictionary will read each word as the corresponding Roman character. If you register with, it will take 2 to 6 bytes for each kanji, occupy a wasted memory space, and since the Roman character strings will be the comparison target when searching the dictionary, it will take more than double the time required and it will correspond to each word Since the Roman character string is not fixed length, there is no regularity in the dictionary structure and it is not easy to search.
Further, if it is attempted to identify the above-mentioned similar phoneme elements, first, each Roman character string is separated into corresponding phoneme element units, and complicated processing such as determining whether they are similar by referring to the correspondence table is performed. It must be done, so it is inefficient and impractical.

In view of the above point, the present invention has an object to provide a kanji conversion device that enables a compact dictionary and can correctly convert an ambiguous input into a kanji string with a high probability.

Means for Solving the Problems The present invention divides the phoneme elements of a mother and a phoneme element, which are similar to each other, into groups, and assigns a bit pattern having a distance of 1 between the phoneme elements of each group. , A dictionary that stores a set of readings of Chinese words represented by using a syllable code that indicates a syllable of one Chinese character by combining with a bit pattern of the vowel and a bit pattern that represents a tone and the corresponding Chinese character code A kanji conversion device provided with dictionary search means for searching the dictionary by masking unclear bit positions of the corresponding syllable code for ambiguous phonetic notations.

Action The present invention makes the dictionary compact by the above configuration,
And by having a regular structure, even for ambiguous input, mask certain information in the reading part of the dictionary,
By equating similar phoneme elements, it is possible to convert into a required kanji string.

Embodiment FIG. 1 is a block diagram of a kanji conversion device according to an embodiment of the present invention. Table 3 is a bit pattern assigned to a vowel, a consonant, a main vowel, a combined vowel, and a voice in the syllable code used for the internal processing of the kanji conversion device in the embodiment of the present invention. These bit patterns are as follows. Is defined by 2 bytes, and at the same time, the distance between the bit patterns of the phoneme elements similar to each other is 1 (the difference bit is the least significant bit).

The bito of the 1st and 2nd bytes is 0, and the bit pattern of the combined vowel is represented by the combination of the bit patterns of the consonant and the main vowel. The syllable code used in the internal processing of the Kanji conversion device of this embodiment corresponds to the ASCII CODE graphic character, that is, according to the syllable code of this embodiment, an arbitrary Chinese syllable has two ASCII CODE graphic characters. Can be expressed as Further, by masking bit5, bit7 of the first byte and bit7 of the second byte, it is possible to identify the same voice mother, consonant, and main rhyme.

In FIG. 1, 10 is at least a phonetic character and an input means having a dictionary search mode key for designating a dictionary search mode, and 11 is a syllable for converting a phonetic character string sent from the input means into the syllable code. A conversion means, 14 is a dictionary storing a set of readings of Chinese words represented by using the syllable code and Kanji codes corresponding to the words, and 13 is a dictionary search mode sent from the input means. By the specification of
When searching the above dictionary, mask the 1st byte bit5, bit7, the 2nd byte bit7 of each syllable of the word reading registered in the dictionary, and use the same mother and interface as the corresponding input phonetic character string. Identifies sounds and main syllables, bit 1 to b of the 2nd byte
It is a dictionary search means for extracting all relevant word candidates from the dictionary by masking it3 and ignoring the tone, or masking the corresponding bit pattern of each phoneme element and ignoring the phoneme element. 12 is the syllable code conversion means
The syllable code sent from 11 is sent to the dictionary search means 13 for each conversion unit, and the word candidates sent from the dictionary search means 13 are selected by the user and the corresponding word candidates are sent to the output means 15. It is a conversion means.

The operation of the Kanji conversion apparatus of this embodiment configured as described above will be described below.

First, the phonetic character input from the input means 10 is converted by the syllable code conversion means 11 into syllable codes for each syllable according to the third table. For example, Is entered, A syllable code such as ASCII CODE graphic ch
When expressed by aracter, it becomes “v ″ h +”. In the syllable code conversion means 11, the input phonetic For each syllable of a character string, it is only necessary to assign the corresponding bit pattern shown in Table 3 according to the phonological element and the tone contained in it, so the conversion is very simple. After the input phonetic character string is converted into a syllable code, the dictionary search means 13 then extracts all applicable word candidates from the dictionary 14 in the dictionary search mode specified by the input means 10. In the above example, the dictionary search key is “v ″ h +”
Only 4 characters of the conventional kanji-kanji conversion device Compared with 10 characters, the number of comparison characters required for search is less than half, and the search speed is faster than before. Besides, since the syllable code is used in the dictionary 14, the amount of memory required to represent the reading of each word is directly proportional to the number of characters that make up that word,
As shown below, when classified according to the number of characters that make up a word, it is possible to have a regular structure with a smaller amount of memory than a conventional dictionary.

In the above example, when the user wants to input "student", the reading is If you have no idea, If you enter the wrong reading like "," it cannot be converted correctly as "student". However, according to the syllable code of this embodiment, The syllable of is represented as follows.

Therefore, by simply pressing the key provided on the input means 10 to identify similar phonological elements, the first byte bit5, bit7, and second byte bit7 of each syllable are masked during the dictionary search, and the "student" To be detected. Furthermore, if you press the key provided to ignore the tone, Is detected. At this time, the kanji conversion means 12 converts the word into a desired word according to the user's selection. Of course, it is possible to easily notify the user of the reading by converting the syllable code registered in the dictionary 14 into the designated phonetic system by a certain routine. This is because the positions of the bits corresponding to the syllable code, the syllable, the consonant, the main syllable, the combined syllable, and the syllable are fixed, so that the conversion can be easily performed only by referring to the correspondence table.

The present invention is not limited to the above-described embodiments, but can be modified and implemented as appropriate without departing from the scope of the invention. For example, the input means 10 is not limited to the input of phonetic character strings by the keyboard,
The input of the voice signal may be changed to an input means for generating a corresponding phonetic character string by voice recognition. Syllable code conversion means
As shown in Table 3, the phonetic system of the phoneme elements in the phonological element-bit pattern correspondence table used in 11 is not only the phonetic system of the above-mentioned vowel sound 1 but also various phonetic systems of the phoneme elements at the same time. By preparing a phonetic character string corresponding to, it is not necessary to change the expansion of the dictionary or the like in the above embodiment by simply switching the input phonetic method, and at the same time, input character strings of various phonetic methods can be used. Can respond. In both cases, since the syllable code is used for internal processing, a simple syllable code converting means and the reading of words in the dictionary are represented by syllable codes, so that input character strings of various phonetic systems can be easily dealt with.

The dictionary search mode of the dictionary search means 13 is not limited to the search mode in which all the corresponding words are extracted from the word dictionary by masking the specific bits and identifying similar phoneme elements. By special designation,
At the time of dictionary search, it is also possible to search by masking a specific part of the input syllable with reading. For example, let “*” be a symbol representing an unknown part of the syllable in the input string. Then, Is input, the bits of the syllable code corresponding to the first syllable and the bits corresponding to the main syllables are masked, and all word candidates satisfying the condition are searched. Such a search is easy because the bits corresponding to each phoneme element of the syllable code are fixed.

Note that the kanji conversion means 12 can mask not only a word-based conversion, but also a kanji conversion device that uses a free sentence as a conversion unit to mask a specific position of a reading part of a dictionary using a syllable code according to the present invention. It is also applicable to select a word candidate to be converted by masking an ambiguous phonetic notation so as not to perform matching on an uncertain bit position of a corresponding syllable code.

The range of application of syllable codes is not limited to Kanji conversion devices,
It can also be used for arbitrary processing of Chinese reading, storage in memory, and transfer between computers.

As described above, according to the present invention, by using the syllable code, the memory capacity of the dictionary is reduced to 80% or less of the conventional one, and at the same time, it is possible to support input by various phonetic methods and By masking the bits in the position of
Even for ambiguous inputs, you can select the necessary word candidates and convert them. In addition, the amount of memory required to register the reading of each word and the kanji code corresponding to that reading is directly proportional to the number of characters that make up the word, so the dictionary can be easily ROMized and the dictionary is economical. The search speed and kanji conversion speed have also been improved, and their practical effects are great.

[Brief description of drawings]

FIG. 1 is a block diagram of a Kanji conversion apparatus according to an embodiment of the present invention, and FIG. 2 is a block diagram of a conventional Kanji conversion apparatus. 10 ... input means, 11 ... syllable code conversion means, 12 ... kanji conversion means, 13 ... dictionary search means, 14 ... dictionary, 15 ... output means.

Claims (1)

[Claims] 1. For a plurality of groups formed by classifying Chinese phonological elements, that is, a mother and a mother, which are similar to each other, a bit having a distance of 1 between constituent elements in each group. Allocating a pattern, reading the Chinese word represented by using a syllable code created so as to show one syllable of a Chinese character by a combination of the bit pattern of the above-mentioned mother, the mother and the bit pattern representing the tone and the above word And a dictionary storing means for storing a set of kanji codes corresponding to, and a dictionary search means for masking an unclear bit position of the corresponding syllable code for an ambiguous phonetic notation. Characteristic Kanji conversion device.