JP3234338B2 - Kanji conversion device - Google Patents

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 converting a phonetic character string into kanji.

[0002]

2. Description of the Related Art There are more than 10,000 Chinese characters used in Chinese sentences. The most important issue in Chinese computer processing including a Chinese word processor is how to correctly and quickly input a kanji intended by a Chinese text creator or the like and to convert it into kanji. Next, as input means to the conventional intended kanji conversion device, there are voice recognition, character recognition, a keyboard and the like. Of these, keyboard input is the most reliable and is widely used.
The input method of kanji using the keyboard is divided into an input method based on reading kanji and an input method based on the kanji character shape. By the way, in the former input method using the character shape, it is necessary to memorize an input rule in advance, and it takes a considerable time to memorize the input rule, and it takes time to get used to it. On the other hand, the input method by reading kanji is widely used in Japanese word processors, and is considered to be the mainstream of the kanji input method in the future because it is natural and easy to learn. The present invention also relates to a kanji conversion device employing this reading input method.

[0003] As a conventional kanji conversion apparatus using this reading as an input, for example, a republic of China Patent Application No. 7
There is such as shown in 5105838. FIG.
Is a block diagram of the kanji conversion device. In the figure, 1
00 is a sound that the Chinese text creator intends to convert to kanji,
This is an input unit for inputting phonetic characters such as note sounds and Roman characters, and is capable of inputting an arbitrary length (number of phonetic characters). 18
Reference numeral 0 denotes a dictionary unit in which a phonetic character string is registered (permanently stored) in association with a word to be converted correspondingly. Then, when a plurality of Chinese words correspond to the same phonetic character string, the total number of strokes of the Chinese word, the frequency of use, and when it is composed of one Chinese character, the frequency of use of the Chinese character, the order of character codes, etc. Are arranged according to the order, and are converted with priority in this order. Of course, if the conversion result is not intended by the input person, the next word or kanji is output by a separate operation in the same manner as in the Japanese word processor. Note that a phonetic character string is given a “string” because, in principle, a plurality of phonetic characters are integrally converted to a single word or kanji. This is a concept that also includes phonetic characters. This is the same for syllable strings and the like that appear later. In addition, the word includes one kanji, and it is a matter of course that the word is not necessarily limited to a Chinese word such as “Japan” or “Tokyo”. Reference numeral 140 denotes an NCHAR register that stores the number of syllables of the input phonetic character string.
Reference numerals 120 and 130 denote a PTR register and an NP register used when converting a phonetic character string into a word, respectively. The PTR register 120 stores from which position of the input phonetic character string the kanji conversion is performed. , NP register stores the converted word length when converting an input phonogram string into a word, that is, the number of kanji or syllables constituting a word (in Chinese, one kanji character is in principle one syllable). I do. 150, after all the conversion processing of words having a certain length or the number of kanjis constituting them,
This is a comparison unit that operates the conversion control unit so that by reducing the value of the P register 130 by one, the kanji conversion is performed with priority given to the next word having one less kanji.
Reference numeral 160 indicates the number of kanji or the number of syllables constituting the word to be converted set in the NP register 130 by sequentially shifting the setting position of the PTR register 120 backward from the first position of the input phonetic character string. It checks whether there are any syllables that have been converted to kanji, and if none have been converted, and if the corresponding word has been registered in the dictionary section 180, it acts to convert it to the corresponding word. This is a conversion control unit. Reference numeral 170 denotes a dictionary search unit that searches the dictionary unit 180 using the syllable string sent from the conversion control unit 160 as a key. Reference numeral 190 denotes an output unit that outputs a result converted by the conversion control unit 160.

With the above configuration and operation, the Chinese word corresponding to the phonetic character string input by the input unit is the longest match method, that is, the first is to give priority to the number of constituent Chinese characters of the word, and the second is to input first. The displayed syllables are converted to kanji in a manner that gives priority to the syllables, and a Chinese sentence is eventually created.

[0005]

However, in the kanji conversion apparatus as described above, words registered in the dictionary unit are used as conversion units and converted by the longest match method.
Therefore, the length of candidate words (the number of kanji that composes them) is the same,
In addition, when a certain syllable in the input phonetic character string forms two corresponding words (a preceding word and a following word) with the preceding and following syllables, the word corresponding to the earlier input syllable is used. (Prior word) is converted preferentially, and after converting the word, the syllable corresponding to the following word is converted to a syllable excluding the syllable common to the preceding word, and in many cases, one kanji character as a unit. Become. In this case, since kanji conversion is performed for the remaining syllables, the following erroneous conversion occurs in the arrangement of kanji for the same syllable in the dictionary. For example, when trying to enter "Aichiten",
There is a kanji character "mass" with the same pronunciation as "ichi". Therefore, the phonetic character string corresponding to the input "yuichiten" has words before and after "useful" and "ichiten". For this reason, the previous word “useful” is preferentially converted, and thereafter “heaven” is converted, so that an erroneous conversion of “useful sky” occurs. Also, when trying to output "murder criminality", there is a kanji "kan" with the same pronunciation as "crime", and a kanji "in" with the same pronunciation as "ze". Since the word "exemplary" is composed, there will be words before and after "exemplary" and "crime", so the prior word "exemplary" will be preferentially converted, and the same pronunciation as the kanji "sin" Is arranged in such a way that it is preferentially converted over "sin" in addition to the presence of the kanji character "". In addition, the Chinese sentence "That characteristic" can be converted to the wrong "odd special expedition". On the other hand, it is difficult as a matter of fact to register all Chinese sentences which may cause such conversion in a dictionary. Therefore, it has been desired to realize a kanji conversion device capable of preventing such erroneous conversion. The present invention has been made for the purpose of solving such a problem.

[0006]

In order to achieve the above object, the present invention first converts a not-yet-converted syllable or part of an input phonetic character string into a word having a maximum number of characters. Second, giving priority to converting the syllable that was input first. Under this condition, the number of syllables to be converted is sequentially reduced, and the syllable to be converted is sequentially shifted backward. A syllable extraction unit that extracts a syllable string to be converted, a dictionary unit in which a phonogram character string and a corresponding Chinese word are registered in advance, and a syllable to be converted extracted by the syllable extraction unit a column as a search key, and the dictionary search unit to search the dictionary portion to detect a word of Chinese, the reading of a special collocation character and該文
In the syllable character dictionary that stores the characters, and in the syllable string to be converted cut out by the syllable cutout unit , the syllable character dictionary is used to determine whether the first syllable is a word.
A concatenated character detecting unit for searching and detecting the concatenated character, and when the corresponding concatenated character is detected by the concatenated character detecting unit, the syllable from the second syllable string of the syllable string to be currently converted is used as a search key, An extended word generation unit that searches the dictionary unit by a search unit, and if there is a corresponding word, generates an extended word having a larger number of constituent Chinese characters by combining the concatenated character and the word. Extended words,
A kanji conversion device comprising: a conversion processing unit that controls to convert kanji preferentially to the word searched by the dictionary search unit.

[0007]

According to the present invention, the syllable cut-out unit gives priority to first converting a part of a syllable character string that has not been converted or a part of the syllable into a word having the maximum number of characters. Second, the number of syllables to be converted is sequentially reduced while giving priority to converting the syllable input first,
Further, the syllables are sequentially shifted backward to cut out a syllable string to be converted into the current kanji. Phonetic character strings and corresponding Chinese words are registered in the dictionary section in advance. The dictionary search unit searches the dictionary unit using the cut syllable string as a search key to search for a corresponding Chinese word. The bilingual character dictionary is
It stores the reading of special collocation characters and the characters. In the syllable string collocations shaped detector is converted cut out by the syllable clipping unit one second syllable is a complex word shape
The collocation character dictionary is searched to determine whether the collocation character is present or not . When the extended word generation unit detects the corresponding collocation character by the collocation character detection unit, a search is performed in the dictionary unit by the dictionary search unit using the other syllables from the second of the syllable string to be currently converted as search keys. If there is a corresponding word, the combination word and the word are combined to generate a longer expanded word. A conversion processing unit converts the expanded words generated by the expanded word generation unit into kanji characters prior to the words searched by the dictionary search unit.

[0008]

DESCRIPTION OF THE PREFERRED EMBODIMENTS The present invention will be described below based on embodiments.
FIG. 1 is a configuration diagram of an embodiment of a kanji conversion device according to the present invention. FIGS. 2 to 4 are processing flowcharts of this embodiment. In FIG. 1, reference numeral 10 denotes an input unit for inputting phonograms such as kanji intended for conversion by a Chinese sentence creator, shouting sounds, note sounds, and Roman letters corresponding to sentences. Reference numeral 18 denotes a dictionary unit in which a phonetic character string and a corresponding Chinese word are registered (permanently stored in advance). When a plurality of words exist in one phonetic character string, these words are arranged in descending order of the frequency of use, and thereby the kanji conversion is prioritized. Reference numeral 14 denotes an NCHAR register indicating the number of syllables in the input phonetic character string. In Chinese, one kanji corresponds to one syllable, so the number of syllables of the input phonogram string is the number of kanji to be converted. 1
Reference numerals 2 and 13 denote a PTR register and an NP register used when converting a phonetic character string into a word, respectively. Here, the PTR register 12 stores a conversion position from which position of the input phonogram character string is to be subjected to kanji conversion. N
The P register 13 stores a word length to be converted when the input phonogram string is converted into a word, that is, the number of kanji and syllables constituting the word. That is, assuming that the contents of the two registers are ptr and np, respectively, it indicates that np syllables consecutive from the ptr-th of the input phonogram string are search keys at the time of dictionary search ( Hereinafter, nchar, np, ptr and nch which will be described later
Let ar represent the values of the NP, PTR and NCHAR registers, respectively). Reference numeral 11 denotes a register initialization unit. Although the details of the detailed operation will be described later, the number of syllables nchar of the input phonogram string is calculated, the value is set in the NCHAR register 14, and if the value is registered in the dictionary unit 18. If the longest word, that is, the number of kanji characters of the word having the largest number of kanji characters is larger than max (8 in this embodiment), the above-mentioned max is set to
The syllable number of the input phonogram string is set in the NP register 13 and the value of the PTR register 12 is set to 1 if the number is 8 or less. Fifteen
Is, (np + ptr)> ( nchar + 1) is constantly inspect whether formed elevational, if satisfied if the value of the NP register 13 reduces 1, the value of the PTR register 12 again 1
The comparison unit is set to. As a result, a word having one less kanji is configured as a target of kanji conversion, and a word having the target number of characters from the beginning of the input phonogram is converted to kanji. Reference numeral 21 denotes a combination word character dictionary which stores special word combinations and stores the character. 20
Is a collocation character detection unit that searches the collocation character dictionary unit 21 using the ptr-th syllable of the input phonetic character string as a search key. 19, when the presence of a corresponding collocation character is detected by the collocation character detection unit 20, the dictionary retrieval unit 17 uses the np-1 syllables from the ptr + 1th of the input phonogram character string as a retrieval key, The dictionary unit 18 searches whether or not the corresponding word is registered. If the corresponding word is registered, the input person intends to perform the conversion by combining the corresponding word with the concatenated character. This is an extended word generation unit that generates a word that is being executed, that is, an extended word. Reference numeral 16 denotes a conversion processing unit. The detailed operation will be described later. Of the ntr syllables consecutive from the ptr-th syllable of the input phonetic character string, the presence or absence of the already converted syllable is checked. The value of PTR register 12 is set to 1
If there are no converted syllables, and if
If there is a corresponding word, it is converted to kanji,
np is added to ptr, and if there is no corresponding word, 1 is added to ptr to process the syllable character string of the next lower rank. Reference numeral 17 designates a word corresponding from the dictionary unit 18 as a search key using the syllable sent from the conversion processing unit 16 or the expanded word generation unit 19, and if there are a plurality of words, the word is arranged first. The most probable word is taken out and converted to the conversion processing unit 16 or the extended word generation unit 1
9 is a dictionary search unit to be sent. 22 is the conversion processing unit 1
6 is an output unit that outputs the result of the conversion.

The operation flow of the embodiment will be described below.
Note that FIGS. 2 to 4 are originally one figure, but are divided into three figures due to the limitation of the screen size. FIG.
Is mainly for the flow of the operation of the initialization of the register, and this operation will be described with reference to FIG. (S1) First, the register for calculating the number of syllables is set to zero.

(S2) A phonetic character string is input. (S3) It is checked whether or not the currently input phonogram is a tone key. If it is a tone key, go to (S4). If not, go to (S5). (S4) With the increase in the number of syllables input, the contents of the register for calculating the number of syllables are increased by 1 and the process returns to (S2).

(S5) It is checked whether or not the currently input phonogram is an input end key. If it is an input end key, go to (S6) to convert to kanji. If it is not an input key (S2)
Return to (S6) It is checked whether the number of syllables currently input exceeds eight. If it is 9 or more, go to (S7). If it is less than 8, go to (S8).

(S7) The value of the NP register is set to 8. (S8) The value of the NP register is set to the currently input syllable number. (S9) The value of the NCHAR register is set to the currently input syllable number. (S10) The value of the PTR register is set to 1. As a result, the initial setting of processing from the head of the input phonogram character string as a target of kanji conversion from a word having a large number of syllables and a large number of kanji characters has been completed.

Next, FIGS. 3 and 4 focus on the operation of the conversion processing unit, and this operation will be described with reference to FIG. (S11) Np syllables that are continuous from the ptr-th syllable character string are cut out. (S12) It is checked whether or not the syllable extracted in (S11) has a converted syllable. If there are already converted syllables, go to (S16). At this time, syllables that have not yet been converted are subjected to conversion when the value of np becomes smaller. If there is no syllable already converted, go to (S13).

(S13) It is searched whether or not the word corresponding to the syllable cut out in (S11) is stored in the dictionary section.
If there is, go to (S14), otherwise go to (S17). (S14) Convert np syllables consecutive from the ptr-th of the input phonetic character string to the corresponding words detected in (S13),
Then go to (S15). (S15) Add np to the value of the PTR register. This is (S
This is because the syllable character string immediately after the kanji conversion in step 14) is to be processed by the next kanji conversion.

(S16) 1 is added to the value of the PTR register.
This is because the next syllable character string having the same number of syllables is to be processed for kanji conversion. (S17) The current value of the NP register is compared with 2. If it is larger than 2, go to (S18); if smaller, go to (S16). (S18) The first syllable cut out in (S11) is checked by a predetermined procedure whether or not it is a reading of a collocation character. if,
If it is a reading of collocation characters, go to (S19), otherwise go to (S16).

(S19) Of the syllables cut out in (S11), it is checked whether or not words corresponding to np-1 syllables from the second syllable are registered in the dictionary unit. If it is registered, go to (S20). If not, go to (S16). (S20) The collocation characters detected in (S18) and the word extracted in (S19) are combined into an extended word, which is one word, to form np syllables consecutive from the input syllable ptr. To an expanded word. Then go to (S15).

In (S21), the value of (np + ptr) and (np + ptr)
Check the magnitude of the value of (char + 1). If the former is larger, go to (S22). If smaller, go to (S11). (S22) Decrement the value of the NP register 13 by 1, set the value of the PTR register 12 to 1, and go to (S23). this is,
This is because a word having one less kanji character is processed as a target of the next kanji conversion from the beginning in the input syllable character string.

In (S23), it is checked whether or not the value of the NP register 13 is zero. If it is zero, the kanji conversion processing of the input phonetic character string ends. Otherwise (S11)
Return to Next, the collocation character dictionary section according to the main part of the present invention will be described. FIG. 5 is a conceptual diagram of a data structure of the collocation character dictionary part of the present embodiment. This script consists of the readings of the basic digraphs and the corresponding digraphs. In the present embodiment, the prepositions "Present", "Subordinate", "Sho", etc. in Chinese, and conjuncts "Rule", "Waka", "Wa", "I" and the like, as well as the adjectives "No" , "Not", "existing", etc., and similarly the designators "its", "the" etc. are regarded as collocation characters. Note that in Chinese, these collocation characters are connected to words to form words having a larger number of kanji characters. Note that a word formed by such a combination is referred to as an extended word in this specification. For example, “John 3 years old”, “Taipei”, “Unnecessary”,
"Yuichiten" and so on. These extended words are regarded as one word (an idiom) in this embodiment. And
In the kanji conversion device adopting the longest match method, conversion is performed with higher priority.

Next, the present embodiment constructed as described above will be specifically described using words as examples. (Less than,
Due to limitations on the characters that can be used by filing through the electronic information processing system, in this specification, FIG.
The phonetic characters and symbols shown in (b), (c), (d), (e), and (f) are inventions intended for the Chinese word processor and the like even in the regulations for the electronic information processing system. For this reason, it is presumed that it was not assumed that a large number of external characters were used in a complicated manner, and similar “sha”, “ren /”,
"Shi @", "fan @", "zui @", "de"
・ ". ) “Sha ren / shi＼ fan＼ zui＼
The operation in the input string "de." will be described. The following shows all the words that can be converted and stored in the dictionary corresponding to this input string.

"Sha ren / murder" "ren / shi @ personnel" "fan @ zui example" "zui de crime" The explanation will be made again based on the flow chart.

When the above phonogram string is input (S2), the register initializing section obtains the number of syllables by dividing the input phonogram string into syllables using tone signals (S3 to S3).
Five) . Since the above input string has six syllables, the number of syllables is set in the NCHAR register (S9). Since the value is smaller than the number of kanji characters 8 (8 in this embodiment) of the longest word in the dictionary unit (S6), 6 is set as the initial value of the NP register (S8), and the value of the PTR register is set to 1 Set to
(S10). Here, the role of the NP register is to store the number of Chinese characters of the word to be obtained by the current conversion. In this case, since 6 is input by default, the kanji conversion is performed in order from the word having 6 syllables. The value ptr of the PTR register indicates that the current dictionary search start position is the ptr-th syllable of the input phonogram string. In this case, since 1 is input by default, the first syllable of the phonogram character string is sequentially converted to kanji.

After the initial values of the PTR register, the NP register, and the NCHAR register are set based on the number of syllables input, the conversion processing unit firstly sets the PTR register and the NP register.
Depending on the value of the register, the first to sixth consecutive syllable strings “sha ren /
shi＼ fan＼ zui＼ de.
(S11) Then, it is checked whether there are any syllables which have already been converted into kanji (S12). If all syllables have not been converted yet, and since these syllable strings have just been input and none of them have been converted to Kanji yet, send all six syllables to the dictionary search unit as search keys. The dictionary is searched (S13). Since there is no corresponding word and the value of the NP register is currently greater than 2 (S17), it is checked whether the first syllable is a reading of a collocation character (S18). If it is determined that the first syllable is not a collocation character, only increase the value of the PTR register by 1 (S
16) At this time, the comparison unit sets the value of (np + ptr) to (nc)
(har + 1), it is determined that this is the case (S21), indicating that another syllable string cannot be cut out to convert a word having six syllables. Therefore, the value of the NP register is decremented by 1 to 5 and PT
The value of the R register is set to 1 again (S22). Since the value of the NP register is not zero (S23), the operation for cutting out the next continuous syllable and processing the kanji conversion is started. Similarly, “sha ren / s” first cut out by the new value 1 of the PTR register and the value 5 of the NP register
hi {fan} zui} ", and since this is also not converted to kanji, prt is added to 1 to 2 (S16), and then the extracted" ren / shi @ fa "
(n＼ zui＼ de.) (S11) is a syllable that has not been converted yet (S12), but the word corresponding to the dictionary part is not stored (S13). (S18), the kanji conversion process is not performed. However, only the value of the PTR register is increased by 1 (S16).
“Ren / shi＼ fan＼ zui＼ de.”
Is determined not to be the object of the kanji conversion (S13), the value of the PTR register is further increased by 1 to become 3. In this case, the comparison unit (np + ptr)> (nchar +
It is determined that 1) (S21), and it is determined that it is no longer possible to cut out another syllable string at the rear in order to convert a word having 5 kanji characters. Therefore, the value of the NP register is reduced by 1 to 4. Further, the value of the PTR register is newly set to 1 (S22). Since the value of the NP register is not zero (S23), the next continuous syllable is cut out and the kanji conversion process is started. When the value of the NP register is 4, as in the case of 5, the continuous syllable string “sh” composed of four pieces sequentially cut out from the input phonogram string by the values of the PTR register and the NP register.
a ren / shi {fan} "," ren / s
Since “hi＼fan＼zui＼” (S11) has not been converted yet (S12), the dictionary is searched using these as search keys. There is no corresponding word (S13),
And because the first syllable is not a reading of collocations (S18),
These are not subject to kanji conversion processing. At this stage, P
The value 2 of the TR register is further increased by 1 (S16). PTR
When the register value becomes 3, the third phonetic character string “sh
i＼ fan＼ zui＼ de. '
(S11), this is also a syllable that has not yet been converted (S12)
. By the way, at this stage, the word corresponding to this syllable string is not stored in the dictionary part (S13), and the value of the NP register is larger than 2 (S17). A check is made to determine if the answer is NO (S18). Then, since the reading "shi @" is registered in the concatenated character dictionary, it is searched whether or not the corresponding word is stored in the dictionary for the other syllables "fan @ zui @ de." (S19). . However, since the corresponding word is not registered, the generation of the expanded word and the kanji conversion are not performed. Next, the value of the PTR register is increased by 1 to 4
(S16). At this time, (np + ptr)> (n
char + 1) (S21), and it is determined that it is no longer possible to cut out another syllable string at the back in order to convert a word having four syllables. Therefore, the value of the NP register is reduced by 1 to 3. Further, the value of the PTR register is set to 1 again (S22). In this case, since the value of the NP register is not zero (S23), the next continuous syllable is cut out and converted.

Even when the value of the NP register is 3, 6
Similarly to the case of 5 and 4, the syllable strings “sha ren / shi＼”, “ren” continuous from the input phonetic character string are set by the action and the set value of the PTR register and the NP register.
/ Shi {fan} "is sequentially cut out (S11).
In this case, all of these syllables have not yet been converted (S
12), they are searched in the dictionary using these as search keys, but there is no word corresponding to them (S13), and the first syllable is not a reading of collocation characters (S18). It will not be done. However, the value of the PTR register is sequentially increased by one (S16). "Shi @ f" cut out when the value of the PTR register becomes 3
an＼ zui＼ ”(S11) is not yet converted (S11
12) Also, since there is no corresponding word in the dictionary section (S13) and the value of the NP register is larger than 2 (S17), it is checked whether the first syllable is a reading of a collocation character (S17). S1
8). By the way, in this case, there is a character "!" Corresponding to the reading "shi @" in the collocation character dictionary. For this reason, it is searched whether or not the corresponding word is registered in the dictionary part for the other syllable "fan {zui}" (S19).
. Since the corresponding word “crime” has been registered, an extended word “correct crime” is generated by combining the collocation word “「 ”with the corresponding word“ crime ”(S20). And "s
hi {fan} zui} "is converted into the expanded word" criminal offense ". Next, the value of the PTR register and the value of the NP register are added. As a result, the value of ptr becomes 6 (S15).
For this reason, (np + ptr)> (ncha
r + 1) (S21). As a result, the number of syllables becomes 3
It is determined that it is not possible to cut out another syllable string at the rear in order to perform the word conversion, and the value of the NP register is reduced by 1 to 2. Further, the value of the PTR register is set to 1 again (S22). In this case, since the value of the NP register is not zero (S23), the next consecutive syllable is cut out and the kanji conversion process is performed.

When the value of the NP register is 2 and the value of the PTR register is 1, the continuous syllable string "sha
ren / "is cut out (S11). Since the syllable is not yet converted (S12), the dictionary search unit searches the dictionary unit using the syllable as a search key. Since the corresponding word "murder" is stored in the dictionary section, "sha ren /"
Is converted to "murder", and the value of the NP register is added to the value of the PTR register to obtain 3 (S15). At this time, the comparison unit determines that (np + ptr) is smaller than (nchar + 1) (S21). Since it is possible to cut out another syllable string in order to convert a word having two kanji characters,
Continuous syllable strings "shi @ fan @", "fa
n＼ zui＼ ”and“ zui＼ de · ”(S
11). However, since these syllables include already converted syllables (S12), no processing is performed.

Next, when the value of the NP register becomes 1, regardless of whether the value of the PTR register is 1, 2, 3, 4 or 5, all of the cut syllables have been converted. As a result, no processing is performed. When the value of the PTR register becomes 6, since "de." Cut out at this stage has not been converted yet, it is sent to the dictionary search unit to search the dictionary unit. In Chinese sentences, the use of "target" is the highest among kanji corresponding to the sound of "de.", And since it is arranged first, it is converted to kanji first.
"De." Is converted to "target". Then, the value of the NP register is added to the value of the PTR register to obtain 7 (S15).
. In this case, (np + ptr) becomes (nc)
(har + 1) is determined (S21). Therefore, the value of the NP register is reduced by 1 and becomes 0. At this point, the conversion processing unit determines that the kanji conversion processing of the input phonogram string has been completed, and sends a conversion result of “murder criminality” to the output unit.

As described above, the present invention has been described based on the embodiments. However, it goes without saying that the present invention is not limited to the above embodiments. That is, for example, it is not necessary to start converting all the phonetic character strings of the input phonetic character strings to kanji after the input end key is pressed, and when the tone key is input, that is, one syllable, Every time is input, the conversion from the input end position of the previously input phonogram string to the currently input phonogram string may be performed. Instead of memorizing the collocation characters independently in the collocation dictionary, these flags will be displayed using flags.
At the same time, it may be stored in the dictionary unit corresponding to this flag. For convenience of manufacture or the like, each constituent part shown in the claims may be physically divided into a plurality of parts, or a plurality of constituent parts may be physically formed as one object. They may be combined. For example, part of the processing of the PTR register, the NP register, the comparison unit, and the conversion processing unit 16 in the embodiment corresponds to the processing of the syllable extraction unit. A phonetic character is a concept that includes not only the illustrated phonemic characters and syllabic characters but also phonemic symbols, and if it is subjected to kanji conversion, it is a character such as Japanese kana, Korean Hangul, etc. Is also good. The uses of course include a word processor for Chinese sentences in Japanese. Needless to say, a learning function or the like may be added.

[0027]

As described above, according to the present invention,
When converting the input syllable string to Chinese characters (Kanji), the syllable to be converted should include Chinese prepositions, conjunctions,
For collocations that have characteristics such as affirmatives, negatives, and nouns, there is a reading of that character, and in subsequent syllables, words with a large number of kanji are integrated with the collocations. If there is a word to be created, the combined word and the word are combined to generate an expanded word having a larger number of kanji and a larger number of syllables to be converted. By the way, since the kanji conversion is performed preferentially from the words with a large number of kanji composed on it, it was difficult to simply convert the kanji with the number of kanjis and the preceding word with priority in the prior art. Misconversions such as "murder and criminal" and "its characteristics" can be easily resolved. For this reason, the accuracy rate at the time of kanji conversion in Chinese sentence creation increases, and the technical effect is very large.

[Brief description of the drawings]

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

FIG. 2 is a flowchart of an operation of a register in an initial stage in the embodiment.

FIG. 3 is a flowchart focusing on the operation of a conversion processing unit in the embodiment.

FIG. 4 is a flowchart focusing on the operation of a conversion processing unit in the embodiment.

FIG. 5 is a conceptual diagram of a data structure of a collocation character dictionary in the embodiment.

FIG. 6 is a configuration diagram of a conventional kanji conversion device.

FIG. 7 is a list of phonetic character strings used in the description of the embodiments in the specification.

[Explanation of symbols]

 Reference Signs List 10 Input unit 11 Register initialization unit 12 PTR register 13 NP register 14 NCHAR register 15 Comparison unit 16 Conversion processing unit 17 Dictionary search unit 18 Dictionary unit 19 Extended word generation unit 20 Concatenated character detection unit 21 Concatenated character dictionary unit 22 Output unit

──────────────────────────────────────────────────の Continuing on the front page (72) Inventor Xie Ming-dan Taiwan Thailand Pay-si-ta-an-ch- 10628 / Len-ai-le-Saint-Toan-136-Hao-10-Losun・ Within Sha Tien Chi Chi Shu Kai Far Far Ku Feng Yu Xiang Kong Sue (72) Inventor Keiken Hayashi Taiwan Thailand Pei Shi Tan An Chi・ 10628 ・ Len I Le Saint Tooan 136 ・ Hao ・ 10 ・ Loung Sha Sha Tien Chi Chi Shu Kai Kai Far Khu Feng Yu Xiang Kong Sue (56) References JP-A-63-163570 (JP, A) JP-A-63-136163 (JP, A) (58) Fields investigated (Int. Cl. ⁷ , DB name) G06F 17/21-17 / 28

Claims (1)

(57) [Claims] 1. For a syllable or a part of the syllable that has not been converted yet in the input phonogram string, priority is first given to converting it to a word having the maximum number of characters, and secondly, it is input first. A syllable segmentation unit that prioritizes converting syllables, sequentially reduces the number of syllables to be converted under this condition, and sequentially shifts the syllables to be converted backward to cut out a syllable string to be converted at the present time. A dictionary part in which phonetic character strings and corresponding Chinese words are registered in advance; and a syllable string extracted by the syllable extraction part, which is to be converted, is used as a search key to search the dictionary part for Chinese. A dictionary search unit for detecting words of the type , reading of special collocation characters, and collocation character characters storing the characters.
In the syllable part and the syllable string to be converted cut out by the syllable cutout part, it is determined whether the first syllable is a collocation character.
A collocation character detecting unit that searches the collocation character dictionary and detects the collocation character.When a corresponding collocation character is detected by the collocation character detection unit, the syllable from the second of the syllable string to be currently converted is extracted. As a search key, the dictionary search unit searches the dictionary unit, and if there is a corresponding word, an extended word generation unit that generates an extended word having a larger number of constituent kanjis by combining the concatenated character and the word; A kanji conversion device comprising: a conversion processing unit that controls the expansion word generated by the generation unit to convert the kanji character with priority to the word searched by the dictionary search unit.