JP2001175645A - Chinese input conversion processing device, Chinese input conversion processing method, and recording medium - Google Patents

Abstract

(57) [Summary] [Problem] To automatically correct an erroneous input such as a key typo and obtain a Chinese kanji intended by a user as a conversion result. When a character string to be converted is pin-in input via an input unit, a correction range determining unit determines a range that needs correction in the input character string, and a correction pattern generating unit. Then, a character string within the range is corrected to create a corrected character string candidate effective as a pinyin. Pinyin kanji conversion is performed by the kanji conversion unit 109 using the corrected character string candidates, and the conversion candidates are output. As a result, even if an erroneous input such as a typo is made, a conversion candidate intended by the user can be obtained by the character string corrected.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a Chinese input conversion processing device used for a document creation device such as a word processor or a personal computer capable of inputting Chinese. The present invention relates to an input conversion processing device, a Chinese input conversion processing method used for the device, and a recording medium.

[0002]

2. Description of the Related Art In a document preparation device capable of inputting Chinese characters, phonograms (hereinafter, referred to as Chinese characters) representing Chinese kanji and their readings.
It has a function of inputting Chinese readings in Pinyin, converting them into Chinese characters, and outputting them using a dictionary that associates them with Pinyin). This is called the Pinyin Kanji conversion function (corresponding to the Kana-Kanji conversion function of a Japanese word processor).

In Chinese, the pronunciation of one kanji is always one syllable. One syllable is further composed of a leading consonant called "vowel" (a consonant at the beginning of the syllable) and a portion excluding "vowel" called "final vowel" (including vowels, diphthongs, and nasalized vowels). Have been. In the Pinyin input method, such initials and finals are input as character strings of the alphabet.

When inputting Chinese kanji readings, the Pinyin character string may be mistaken. If the Pinyin string is incorrect, the intended Chinese kanji cannot be obtained. Conventionally, as a method of correcting such an input error, for example, there is a method described in Japanese Patent Application Laid-Open No. Hei 5-108647. This corrects pronunciation errors. That is, for example, similar pronunciations such as “ting” and “ding” are registered in the table in advance, and assuming that the Pinyin character string input by the user is incorrect, A similar character string is output as a correction candidate.

[0005]

As described above, conventionally, there has been a method for correcting an error in pronunciation of an input Pinyin character string. However, such a correction method cannot correct an erroneous input such as a key typo, even if an error with similar pronunciation can be corrected. At the time of inputting Pinyin, not only pronunciation mistakes but also typographical errors are extremely common, and there is a problem that the user has to re-input many times to obtain the intended Chinese kanji.

SUMMARY OF THE INVENTION The present invention has been made in view of the above points, and provides a method of automatically correcting an erroneous input such as a key typo, thereby obtaining a Chinese kanji intended by a user as a conversion result. It is an object to provide a word input conversion processing device, a Chinese input conversion processing method, and a recording medium.

[0007]

According to the present invention, there is provided a Chinese input conversion processing apparatus for inputting a Pinyin which is a Chinese reading notation and for converting the input Pinyin into a Chinese kanji. When the column is pinyin input,
The input character string is provided with input correction means for determining a range requiring correction, and correcting the character string within the range to create a valid correction character string candidate as pinyin. The corrected character string candidates are converted into Chinese kanji, and the conversion candidates are output in a predetermined order.

According to such a configuration, when a part of the input Pinyin character string has an error, the erroneous input part is determined as a correction range, and is valid as Pinyin by deleting, adding, or changing the character. A corrected character string candidate is created. Therefore, even if an erroneous input such as a typo is made, a conversion candidate intended by the user can be obtained by using the character string candidate corrected.

Further, in the above-mentioned configuration, a similar pronunciation detecting means for detecting a pronunciation similar to a character string before correction among candidates of a corrected character string is provided, so that a conversion candidate having a corrected pronunciation error can be detected. Output can be given priority.

Further, in the above-mentioned configuration, the correction learning detecting means for detecting a correction character string candidate in which the same input correction has been performed in the past is provided, so that the conversion candidate intended by the user is given priority. Can be output to

[0011] In the above-mentioned structure, the typo detecting means for detecting, among the candidates of the corrected character string, a key input with a key adjacent to the key corresponding to the character string before correction, the typo detecting means is provided. A conversion candidate in which an erroneous typo has been corrected can be preferentially output from the arrangement of the keys.

[0012]

Embodiments of the present invention will be described below with reference to the drawings.

(First Embodiment) FIG. 1 is a block diagram showing a system configuration of a Chinese input conversion processing device according to the present invention. This apparatus has a function of inputting Pinyin, which is a Chinese reading notation, converting the input Pinyin into Chinese kanji, and outputting it.For example, it is recorded on a recording medium such as a magnetic disk. It is realized by a computer that reads a program and whose operation is controlled by the program.

As shown in FIG. 1, a Chinese input conversion processing device according to this embodiment comprises a CPU 11 and a CPU 1.
A ROM 13 connected to the system 1 via a system bus 12;
It comprises a RAM 14, an input device 15, a display device 16, an external storage device 17, a printing device 18, a communication device 19 and the like.

The CPU 11 controls the entire operation of the present apparatus, and operates by starting a program stored in a memory such as the ROM 13 or the RAM 14. In addition, as a function realized based on the program stored in the memory, there is a word processor or the like by an application program.

The ROM 13 stores data such as various programs and fonts (including Chinese fonts) necessary for starting the apparatus. The RAM 14 stores programs downloaded from the external storage device 17 and the CPU 1.
Various types of data used for the first process are stored.

The input device 15 is composed of, for example, a keyboard and is used to input instructions and data to the present device.
This is for performing a Pinyin Kanji conversion instruction (an instruction by operating a “Conversion” key) for the input Pinyin character string, and the like.

The display device 16 is, for example, a CRT (Cathode
It comprises a Ray Tube) and an LCD (Liquid Crystal Display), and displays a character string input through the input device 15, a conversion result, and the like.

The external storage device 17 is, for example, an HDD (Hard
Disk Drive) and CD-ROM (Compact Disc Read Only)
Memory), which stores data such as various programs and documents. The external storage device 17 includes data such as a dictionary used in Pinyin kanji conversion processing, in addition to an application program for realizing a word processor.

The printing device 18 is used for printing a document or the like. The printing device 18 may be built in the device or connected externally. The communication device 19 is, for example, a local LAN (Local
It controls the transmission and reception of data to and from external devices via an area network) or a communication network.

Next, the functional configuration of the Chinese input conversion processing device realized by the system configuration shown in FIG. 1 will be described.

FIG. 2 is a block diagram showing a functional configuration of the Chinese input conversion processing device according to the first embodiment.

Functionally showing the processing of the CPU 11 in FIG. 1, the present apparatus comprises an input unit 101, a control unit 102, an output unit 103, an input MMI unit 104, a converted character string buffer 10
5, correction pattern generation unit 106, pinyin table 10
7, correction pattern buffer 108, kanji conversion unit 109,
It comprises a conversion dictionary 110, a conversion result buffer 111, a candidate selection unit 112, and a correction range determination unit 113.

The input unit 101 inputs characters and functions. The control unit 102 controls each unit of the input unit 101, the output unit 103, and the input MMI unit 104. The output unit 103 displays the input result. Input MMI
The unit 104 processes a user's input operation.

The conversion character string buffer 105 stores a Pinyin character string input as a conversion target. The correction pattern generation unit 106 generates correction pattern candidates for the Pinyin character string input as a conversion target. The pinyin table 107 stores information for determining whether the character string is valid as pinyin. The correction pattern buffer 108 stores the correction pattern candidates generated by the correction pattern generation unit 106.

The kanji conversion unit 109 targets the input character string stored in the conversion character string buffer 105 and the correction pattern candidates stored in the correction pattern buffer 108 as targets.
With reference to the conversion dictionary 110, it is converted into Chinese kanji. The conversion dictionary 110 stores information necessary for Pinyin kanji conversion processing, such as reading of Pinyin notation and corresponding Chinese kanji for each word.

The conversion result buffer 111 stores the kanji conversion unit 1
09 holds the Chinese kanji candidate obtained as the conversion result. The candidate selection unit 112 performs a candidate selection process according to a user's selection operation. The correction range determining unit 113
Referring to the conversion dictionary 110 or the pinyin table 107, a range that needs to be corrected in the input character string stored in the conversion character string buffer 105 is determined.

Next, the operation of the first embodiment will be described.

FIG. 3 is a flow chart showing the conversion processing operation in the first embodiment.

For example, when any key is input through the input unit 101 composed of a keyboard (step A1)
1) The control unit 102 sends the key input information to the input MMI unit 104. The input MMI unit 104 is a front-end processor (FEP) that performs kana-kanji conversion.
Such functions are assumed, and here, only character input processing directly related to the present invention will be described (the same applies to other embodiments).

Upon receiving the key input information, input MMI section 104 first determines whether or not the key input is a character input (step A12). If the key input is a normal character input, that is, if it is a Pinyin character string input as a conversion target (Yes in step A12), the input MMI unit 104
Stores a character string composed of a character code corresponding to the key input information in the converted character string buffer 105 (step A13).

If the key input is not a character input in step A12, the input MMI unit 104 next determines whether or not a conversion operation has been performed (step A1).
4). In the case of the conversion operation, that is, when Pinyin Kanji conversion is instructed by pressing the “Conversion” key provided in the input unit 101 (Yes in step A14), the Pinyin Kanji conversion processing is usually performed by the Kanji conversion unit 109. Is executed. In the present invention, an input correction process is executed prior to this.

The input correction process is performed in the following procedure.

Now, as a character string to be converted, "Woyeh
It is assumed that a character string such as "nhao" (input error of "wyohhenhao") is pinyin-input by the user.

The correction pattern generator 106 first requests the correction range determiner 113 to determine a correction range.
Upon receiving this request, the correction range determination unit 113 sets the input character string “wo” currently stored in the converted character string buffer 105.
For “ehnnhao”, a range that needs to be corrected in the input character string is determined (step A15). When the entire character string is to be corrected, the number of correction pattern candidates becomes enormous, so correction is performed in advance. It is necessary to determine the necessary range.

The correction range is determined as follows.

The entire range of the input character string is picked up, and the conversion dictionary 110 is searched for the character string in the entire range. As a result, the portion of the character string that can be searched is excluded, and the range of the remaining character string is set as the correction range. Note that instead of searching the conversion dictionary 110, the pinyin table 107 may be searched to exclude a range valid as pinyin. Even if the character string is searched, the character string may not be the intended character string. Therefore, it is preferable to perform the correction process on the character string range searched by the conversion dictionary 110 or the pinyin table 107.

More specifically, a correction range check table as shown in FIG. 10 is created to determine a correction range. The correction range check table shown in FIG.
This shows the result of a dictionary search performed on the partial character string “yehnhao”, where “○” indicates that there was a search result. Or a valid pinyin)
“Wo”, “o”, “ye”, “e”, “n”, “h”
a "," hao "," a "," ao "," o ".

Looking at the input character string "wyohnhao" from the first character, there is a character string that could be searched up to "wo". There is no character string that could be searched for using a character string that includes the second character “o” and the next “y”, and is divided here.
Similarly, the range of “ye” can be searched, but there is no search with a character string including the next “h”, so that there is no search result in this part. The range of “n” and “hao” indicates a search result.

Here, the portion "h" for which there is no search result is the range in which the correction processing is to be performed. However, since there is a possibility that the part before and after this range may be incorrect, it is determined as one correction range including the range before and after that (range 2). Further, the range in which the search result is found may be added as the correction range (range 1, range 3).

The correction range determination unit 113 sends information on the determined correction range (at least the character string within the correction range and the start position on the input character string) to the correction pattern generation unit 106. In this case, the character string in the correction range without the search result (range 2) or in the entire correction range including the range with the search result (range 1 to 3) is the correction pattern generation unit 1.
06.

Next, the correction pattern generation unit 106
A correction pattern is generated for the correction range determined by the correction range determination unit 113 (step A16). Assuming that the character string in the correction range is an erroneous input, it is assumed that extra characters are added to the original character string, necessary characters are missing, and different characters are input. On the contrary, the correction pattern is generated by deleting, adding, and changing characters at each character position. However, considering all of these correction combinations results in an astronomical number combination, the correction is performed here assuming that there is only one error in the input character string for one conversion. In this case, the maximum number of correction patterns in the entire character string
If it is represented by 6 characters, "delete character" is (character length)
This is 26 × (character length + 1) for “addition of characters” and 25 × (character length) for “change of characters”, which are numbers that can be sufficiently processed. If the processing capacity of the apparatus has room, the maximum number of correction points can be increased.

FIG. 11 shows an example of a correction pattern candidate created in this way.

FIGS. 11 (a), 11 (b) and 11 (c) show an input character string "wyehhenhao"("wyehhenha").
FIG. 11 (a) shows a correction pattern candidate in a case where the part of “yehn” in “o” input error) is used as a correction range. In FIG. 11A, one character of “yehn” is deleted. The corrected pattern candidate generated by the above method is shown in FIG.
The correction pattern candidate generated by adding one character between each character of “ehn”, and FIG.
The correction pattern candidates generated by changing to z are shown.

Here, since there are many unnecessary correction patterns as they are, the correction pattern generation unit 106
Performs an effective pinyin check on the created correction pattern candidate (step A17). In this valid pinyin check, it is determined using the pinyin table 107 whether or not the correction portion is valid pinyin. As shown in FIG. 12, the pinyin table 107 stores a character string effective as a syllable represented by pinyin. There are various methods for checking whether the pinyin is valid. One example is a method of determining whether the entire character string is valid as a pinyin (whether continuous pinyin) in order from the front of the character string. There is.

FIG. 13 shows correction pattern candidates determined as valid pinyin as a result of such a check. FIG. 13A shows a correction pattern candidate by deleting characters, FIG. 13B shows a correction pattern candidate by adding characters, and FIG.
Are correction pattern candidates due to character change, and only those determined as valid pinyin are shown. Through such a check, the number of correction pattern candidates can be reduced. The correction pattern determined to be valid pinyin is stored in the correction pattern buffer 108 (step A18).

The correction pattern is stored in the correction pattern buffer 108 in a format as shown in FIG. 14, for example. Here, the number of corrected characters is one for one correction candidate, but when correcting at a plurality of locations, information corresponding to the number of characters is stored. It should be noted that “No” in the figure only indicates the correspondence with the created correction pattern candidate, and therefore it is not actually necessary to store the correction pattern in the correction pattern buffer 108. Also, the “character position” is not the character string divided in the correction range, but the original input character string (“wyohehna”).
o ").

After the correction pattern for the input character string is stored in the correction pattern buffer 108 in this way, the kanji conversion unit 109 performs the Pinyin kanji conversion process with reference to the conversion dictionary 110 (step A19). In this case, the kanji conversion unit 109 performs conversion processing on the input character string in the conversion character string buffer 105, and also performs conversion processing on correction pattern candidates with reference to the correction pattern buffer 108 to create kanji conversion candidates. By this process, conversion candidates are obtained for the corrected pinyin.

FIG. 15 shows an example of kanji conversion candidates.

When a character string such as "wyehhnhao" is input, Pinyin kanji conversion is performed in units of syllables such as "wo / ye / hn / hao"("/" indicates a syllable delimiter). At that time, "wo",
For “ye” and “hao”, conversion processing is performed using a character string without correction, but for “hn”, conversion processing is performed using a corrected character string (“han”, “hen”...). Is performed.

In FIG. 15, the conversion candidates using the corrected character strings are output in the order in which the correction patterns are created. However, in practice, the most appropriate candidates are output with priority. For this, for example, a method is used in which the relationship between words before and after is checked, and those with strong grammatical connection between words are prioritized. The specific rule can be stored in the conversion dictionary 110 or as another logic. FIG.
In the example, “Gayaya” is obtained as the first candidate. The created conversion candidates are stored in the conversion result buffer 11 in the order of priority.
1 is stored.

With the above, the Pinyin Kanji conversion process is completed, and the process returns to the state of waiting for key input (step A11).

If the key input is not a conversion operation in step A14, the input MMI unit 104 next determines whether or not the input operation is a candidate selection operation (step A2).
0). In the case of a candidate selection operation (Yes in step A20)
s) The candidate selection unit 112 performs a candidate selection process (step A21). The candidate selection unit 112 displays the conversion candidates with reference to the conversion result buffer 111, and changes or fixes the display candidates according to the selection operation of the user.

As a method of displaying the conversion candidates, it is general to display the conversion candidates in the order in which they are stored in the conversion result buffer 111.
A selection method such as switching the candidate display to (addition / change) is also conceivable. Furthermore, when there are too many conversion candidates (when there are more than a predetermined number), only the top candidates may be displayed.

Various other methods of displaying the conversion candidates can be considered. However, since the display order of the candidates affects the ease of candidate selection, it is assumed that the candidate with the higher display order is more easily selected.

When a conversion candidate is determined, the candidate selection unit 1
12 returns the determined result to the input MMI unit 104. The input MMI unit 104 returns information on the candidate selection state or the confirmed character string to the control unit 102. The control unit 102 displays a candidate selection state and processes a confirmed character string according to this information, and outputs the processing result to the output unit 103.

As described above, according to the first embodiment, when there is an error in a part of the input Pinyin character string, the erroneously input part is determined as a correction range, and the character deletion,
Correction character string candidates effective as pinyin are created by addition, change, and the like. Therefore, even if an erroneous input such as a typographical error occurs, the conversion candidate intended by the user can be obtained in the conversion result by the Pinyin kanji conversion using the character string candidate corrected.

(Second Embodiment) Next, a second embodiment of the present invention will be described.

The second embodiment is characterized in that a conversion candidate in which a pronunciation error has been corrected is preferentially output.

FIG. 4 is a block diagram showing a functional configuration of a Chinese input conversion processing device according to the second embodiment of the present invention. In FIG. 4, the same portions as those in FIG. 2 (first embodiment) are denoted by the same reference numerals, and description thereof will be omitted.

The basic configuration of this device is the same as that of the first embodiment, except for the equivalent sounding priority section 114,
That is, a pronunciation table 115 is provided. The equivalent pronunciation priority unit 114 refers to the pronunciation table 115, detects a candidate of a corrected character string that is similar in pronunciation to the character string before correction (the original input character string), and detects the corrected character string. Priority is given to conversion candidates obtained based on the candidates. In the pronunciation table 115, a Pinyin character string having a similar pronunciation is registered.

FIG. 16 shows an example of the sound generation table 115.

In the sound generation table 115, "tiao" and "diao", "ting" and "ding", and "pa"
A set of words having similar pronunciations, such as “o” and “lao”, “shang” and “sang”, etc., is registered in Pinyin notation. And those that are prone to mispronunciation, such as "crab" and "crack", and are given the same pronunciation code.

Next, the operation of the second embodiment will be described.

FIG. 5 is a flowchart showing the conversion processing operation in the second embodiment. In FIG. 5,
The processing of steps B11 to B19 is the same as the processing of steps A11 to A19 in FIG.

That is, when a character string to be converted (Chinese reading) is pin-in input from the input unit 101, the control unit 102 stores the input character string in the conversion character string buffer 105 through the input MMI unit 104. Then, it waits for the next instruction (steps B11 to B13). When Pinyin Kanji conversion is instructed by pressing the “Conversion” key, the control unit 102
Executes an erroneous input correction process on the input character string stored in the converted character string buffer 105, and stores the corrected character string candidate (correction pattern candidate) in the correction pattern buffer 108 (steps B14 to B18). Then, the kanji conversion unit 109 performs the Pinyin kanji conversion process with reference to the conversion character string buffer 105 and the correction pattern buffer 108, and stores the conversion candidates obtained as the conversion result in the conversion result buffer 111 (step B19). ).

However, in the second embodiment, the conversion dictionary 1
Correction processing is also performed on the range of character strings that can be searched in 10 or the Pinyin table 107. For example, "woquotiaoyu"("woquotiaoy
u ”), step B15
In the correction range determination process, the input character string is "wo / q
u / tiao / yu ”. In this case, a correction candidate is created for each of the divided character strings.
Only the part of "o" (correct pinyin is "diao") will be described with reference to Fig. 17. Fig. 17 shows the result when the correction pattern candidate for the part of "tiao" is converted into Pinyin kanji.

Here, after the Pinyin kanji conversion, the equivalent sounding priority unit 114 refers to the sounding table 115 and performs output priority processing on the conversion candidates obtained at that time (step B2).
0). As shown in FIG. 18, information (correction pattern No.) indicating the correspondence between the correction pattern and the input character string is added to the conversion candidate with input correction. Note that FIG.
In the table, “head of reading” and “reading length” indicate which range of the input character string corresponds to the candidate. In this example, it is shown that the range corresponds to the range of 4 characters from the (4 + 1) th character of the input character string “woqutiaoyu” (counting the leading character as 0).

The equivalent sounding priority section 114 refers to the conversion character string buffer 105 and the correction pattern buffer 108 to obtain a character string before correction (original input character string) and a character string after correction for each conversion candidate. Then, the pronunciation table 115 is searched with these character strings. As a result, if the phonetic code of the uncorrected character string matches the phonetic code of the corrected character string, that is, if there is a phoneme with similar pronunciation, the conversion candidate obtained based on the corrected character string is used. Move up the ranking. However, it shall not be ranked higher than the candidate whose pinyin has not been corrected.

In the above example, “diao” is replaced by “tia”.
Since the pronunciation is similar to “o”, the corresponding conversion candidate is moved to a higher position, and “diao” and “tiao” have the same pronunciation code added in the pronunciation table 115.

After performing the same processing for all correction candidates, the result is stored in the conversion result buffer 111. As a result, as shown in FIG. 19, the conversion result buffer 111 stores a conversion candidate using a similar sounding correction candidate in response to an erroneous input such as “tiao”. In this example, since there is a conversion candidate without correction such as “tiao”, a conversion candidate of a similar pronunciation “diao” is stored next to the conversion candidate.

The subsequent processing is the same as in the first embodiment.

That is, if the key input is not a conversion operation in step B14, the input MMI unit 104 next determines whether or not the input operation is a candidate selection operation (step B2).
1). In the case of a candidate selection operation (Yes in step B21)
s) The candidate selection unit 112 performs a candidate selection process (step B22). The candidate selection unit 112 displays the conversion candidates with reference to the conversion result buffer 111, and changes or fixes the display candidates according to the selection operation of the user.

The conversion candidates are stored in the conversion result buffer 111.
May be displayed in the order in which they are stored, or conversion candidates without correction and conversion candidates with correction may be switched and displayed.

As described above, according to the second embodiment, among the corrected character string candidates, the one whose pronunciation is similar to the character string before correction is detected, so that the conversion in which the pronunciation error is corrected is performed. Candidates can be output with priority.

(Third Embodiment) Next, a third embodiment of the present invention will be described.

The third embodiment is characterized in that, by learning a correction pattern of an input character string, priority is given to the output of a conversion candidate based on the same corrected character string.

FIG. 6 is a block diagram showing a functional configuration of a Chinese input conversion processing device according to the third embodiment of the present invention. In FIG. 6, the same portions as those in FIG. 2 (first embodiment) are denoted by the same reference numerals, and description thereof will be omitted.

The basic configuration of this device is the same as that of the first embodiment, except that a correction learning priority unit 116, a correction learning registration unit 117, and a correction learning table 118 are provided. The correction learning priority unit 116 performs a process of referring to the correction learning table 118 to prioritize a conversion candidate having the same correction as the learned correction pattern.
The correction learning registration unit 117 performs a process of registering the correction pattern of the conversion candidate selected at the time of conversion in the correction learning table 118. The correction learning table 118 stores the correction patterns registered by the correction learning registration unit 117.

Next, the operation of the third embodiment will be described.

FIG. 7 is a flowchart showing the conversion processing operation in the third embodiment. In FIG. 7,
The processing in steps C11 to C19 is the same as the processing in steps A11 to A19 in FIG.

That is, when a character string to be converted (Chinese reading) is input from the input unit 101, the control unit 102 stores the input character string in the conversion character string buffer 105 through the input MMI unit 104. Then, it waits for the next instruction (steps C11 to C13). When Pinyin Kanji conversion is instructed by pressing the “Conversion” key, the control unit 102
Performs an erroneous input correction process on the input character string stored in the converted character string buffer 105, and stores the corrected character string candidate in the correction pattern buffer 108 (step C).
14-C18). Then, the kanji conversion unit 109 performs the Pinyin kanji conversion process with reference to the conversion character string buffer 105 and the correction pattern buffer 108, and stores the conversion result candidate in the conversion result buffer 111 (step C19). .

Here, in the third embodiment, after the Pinyin kanji conversion, the correction learning priority section 116 performs correction learning priority processing (step C20). Here, the conversion candidate that has been subjected to the correction learning is preferentially output, but since the correction pattern is not registered in the correction learning table 118 at first, the candidate obtained as the conversion result is directly used in the conversion result buffer 1.
Output to 11

Thereafter, when there is a candidate selecting operation (Yes in step C21), the candidate selecting unit 112 performs a candidate selecting process (step C22). Candidate selection unit 112
Displays the conversion candidates with reference to the conversion result buffer 111, and changes or fixes the display candidates according to the user's selection operation. When the conversion candidate is determined, the candidate selection unit 112
Returns the determined result to the input MMI unit 104. Input MM
The I unit 104 returns information on the candidate selection state or the confirmed character string to the control unit 102. The control unit 102 displays a candidate selection state and processes a confirmed character string according to this information, and outputs the processing result to the output unit 103.

Here, when a conversion candidate is selected (confirmed), information on the conversion candidate selected at that time is sent to the correction learning registration unit 117, and correction learning registration processing is performed (step C23). If the selected conversion candidate is a candidate with input correction, a correction pattern corresponding to the conversion candidate is created as shown in FIG. 18 and stored in the correction learning table 118 for learning.

FIG. 20 shows an example of the correction learning table 118.

In the correction learning table 118, characters before correction and characters after correction are associated with each other. In the example of FIG. 20, only one corrected character is stored, but it may be stored including characters before and after the corrected character.
In such a case, the learning priority processing compares the characters with each other, so that the learning effect is distinguished by the immediately preceding character and the immediately following character. In addition, the correction learning table 11
8, it is assumed that the correction patterns are arranged in the order in which they were registered last.

As described above, the correction pattern corresponding to the conversion candidate selected at the time of conversion is stored in the correction learning table 118.
After the registration, the pinyin input is performed again, and the pinyin conversion process is performed (step C19). When the correction learning priority section 116 corrects the correction pattern corresponding to each conversion candidate and the correction pattern stored in the correction learning table 118, Are compared, and when there is the same one, the order of the conversion candidate is moved to a higher rank (step C20). After performing the same processing as above for all the corrected candidates, the result is sent to the conversion result buffer 111. The candidate selection unit 112 performs display / selection of candidates according to the order stored in the conversion result buffer 111.

As described above, according to the third embodiment, the correction pattern of the input character string corresponding to the conversion candidate selected at the time of conversion is learned and the same input character Since the corrected one is detected, the conversion candidate intended by the user can be preferentially output.

(Fourth Embodiment) Next, a fourth embodiment of the present invention will be described.

The fourth embodiment is characterized in that the priority of the conversion candidates based on the corrected character string is determined in consideration of the arrangement of each key on the keyboard.

FIG. 8 is a block diagram showing a functional configuration of a Chinese input conversion processing device according to the fourth embodiment of the present invention. In FIG. 8, the same portions as those in FIG. 2 (first embodiment) are denoted by the same reference numerals, and description thereof will be omitted.

The basic configuration of the present apparatus is the same as that of the first embodiment, and the difference is that the typo correction priority unit 11
9. An input key arrangement table 120 is provided. The typo correction priority unit 119 determines a key adjacent to the input key with reference to the input key arrangement table 120, and performs a process of giving priority to a correction pattern corresponding to the key. The input key arrangement table 120 is a table storing information on keys adjacent to each key on the keyboard.

When the input unit 101 is composed of a keyboard and the keys on the keyboard are arranged as shown in FIG. 21 in accordance with a predetermined standard, the input key arrangement table 120 stores information on each key as shown in FIG. Information on the adjacent key is stored. For example, for the "Q" key,
Information such as “W” and “A” is stored as adjacent keys. For the “A” key, information such as “S”, “Z”, “Q”, and “W” is stored as adjacent keys, and for the “Z” key, “X” is stored as an adjacent key. ,
Information such as “A” and “S” is stored. Similarly, information on adjacent keys for all keys is stored in the input key arrangement table 120.

Next, the operation of the fourth embodiment will be described.

FIG. 9 is a flowchart showing the conversion processing operation in the fourth embodiment. In FIG. 9,
The processing of steps D11 to D19 is the same as the processing of steps A11 to A19 in FIG.

That is, when a character string (Chinese reading) to be converted is input from the input unit 101, the control unit 102 stores the input character string in the conversion character string buffer 105 through the input MMI unit 104. Then, it waits for the next instruction (steps D11 to D13). When Pinyin Kanji conversion is instructed by pressing the “Conversion” key, the control unit 102
Performs an erroneous input correction process on the input character string stored in the converted character string buffer 105, and stores the corrected character string candidate in the correction pattern buffer 108 (step D).
14-D18). Then, the kanji conversion unit 109 performs the Pinyin kanji conversion process with reference to the conversion character string buffer 105 and the correction pattern buffer 108, and stores the conversion result candidate in the conversion result buffer 111 (step D19). .

Here, in the fourth embodiment, after the Pinyin Kanji conversion, the candidates obtained as the conversion result are sent to the typo correction priority unit 119. At this time, it is assumed that the conversion candidate with the input correction stores the correspondence information with the correction pattern as shown in FIG. Typo correction priority section 1
19 performs the following type error correction priority processing (step D20).

First, for each conversion candidate, for the correction pattern whose correction type is “change” of the character, the input key arrangement table 120 is searched with the corrected character. In the case of the key arrangement as shown in FIG. 21, the input key arrangement table 120 stores information on adjacent keys as shown in FIG.

Input key arrangement table 12 with corrected characters
Searching for 0 gives the adjacent key. For example, if the corrected character is “s”, adjacent keys “D”, “X”, “Z”, “A”, “W”, “E” for the “S” key
Is obtained. By comparing these adjacent keys with the character before correction (obtained from the converted character string buffer 105 based on the “character position” stored in the correction pattern), the character after correction and the character before correction can be compared. It is determined whether or not both keys are close to each other. Note that the input key arrangement table 120 may be searched with the character before correction, and may be compared with the character after correction. If it is determined that the conversion candidates are close to each other, the order of the conversion candidates is moved to a higher rank. The same processing as described above is performed for all correction candidates, and the result is sent to the conversion result buffer 111.

The subsequent processing is the same as in the first embodiment.

That is, if the key input is not a conversion operation in step D14, input MMI section 104 next determines whether or not the input operation is a candidate selection operation (step D2).
1). In the case of a candidate selection operation (Ye in step D21)
s) The candidate selection unit 112 performs a candidate selection process (step D22). The candidate selection unit 112 displays the conversion candidates with reference to the conversion result buffer 111, and changes or fixes the display candidates according to the selection operation of the user.

As described above, according to the fourth embodiment, among the candidates for the corrected character string, those input with the keys adjacent to the character string before correction are detected, so that each key on the keyboard is detected. A conversion candidate in which an erroneous typo has been corrected can be preferentially output from the arrangement of.

The methods described in each of the above-described embodiments include, for example, a magnetic disk (floppy disk, hard disk, etc.), an optical disk (CD-ROM,
It is also possible to write the data on a recording medium such as a DVD or a semiconductor memory and apply it to various devices, or to transmit it via a communication medium and apply it to various devices. A computer that realizes the present apparatus reads the program recorded on the recording medium, and executes the above-described processing by controlling the operation of the program.

[0105]

As described above in detail, according to the present invention, an erroneous input portion in an input character string is corrected, and Pinyin kanji conversion is performed using the corrected character string candidate. Even erroneous inputs such as mistakes can be corrected, and conversion candidates intended by the user can be obtained without re-inputting. Therefore, in creating a Chinese document by pinyin input, it is possible to reduce the burden of the user's input operation and realize efficient document creation.

[Brief description of the drawings]

FIG. 1 is a block diagram showing a system configuration of a Chinese input conversion processing device of the present invention.

FIG. 2 is a block diagram illustrating a functional configuration of a Chinese input conversion processing device according to the first embodiment.

FIG. 3 is a flowchart illustrating a conversion processing operation according to the first embodiment.

FIG. 4 is a block diagram showing a functional configuration of a Chinese input conversion processing device according to a second embodiment of the present invention.

FIG. 5 is a flowchart illustrating a conversion processing operation according to the second embodiment.

FIG. 6 is a block diagram showing a functional configuration of a Chinese input conversion processing device according to a third embodiment of the present invention.

FIG. 7 is a flowchart illustrating a conversion processing operation according to the third embodiment.

FIG. 8 is a block diagram showing a functional configuration of a Chinese input conversion processing device according to a fourth embodiment of the present invention.

FIG. 9 is a flowchart illustrating a conversion processing operation according to the fourth embodiment.

FIG. 10 is a diagram showing an example of a correction range check table used when determining a correction range of an input character string.

FIG. 11 is a diagram showing an example of a correction pattern candidate.

FIG. 12 is a diagram showing a configuration of a pinyin table.

FIG. 13 is a view showing an example of a correction pattern candidate after valid pinyin check;

FIG. 14 is a diagram showing an example of storing a correction pattern.

FIG. 15 is a diagram showing an example (part 1) of kanji conversion candidates obtained by erroneous input correction.

FIG. 16 is a diagram showing a configuration of a sounding table.

FIG. 17 is a diagram showing an example (part 2) of kanji conversion candidates obtained by erroneous input correction.

FIG. 18 is a diagram showing a relationship between kanji conversion candidates and correction patterns.

FIG. 19 is a view showing an example (part 3) of kanji conversion candidates obtained by erroneous input correction.

FIG. 20 is a diagram showing a configuration of a correction learning table.

FIG. 21 is a diagram showing an example of the arrangement of keys on a keyboard.

FIG. 22 is a diagram showing a configuration of a key arrangement table.

[Explanation of symbols]

 101 input unit 102 control unit 103 output unit 104 input MMI unit 105 conversion character string buffer 106 correction pattern generation unit 107 pinyin table 108 correction pattern buffer 109 kanji conversion unit 110 conversion dictionary 111 conversion Result buffer 112 ... Candidate selection unit 113 ... Correction range determination unit 114 ... Equivalent pronunciation priority unit 115 ... Speech table 116 ... Correction learning priority unit 117 ... Correction learning registration unit 118 ... Correction learning table 119 ... Type error correction priority unit 120 ... Input Key arrangement table

 ──────────────────────────────────────────────────続 き Continuing from the front page (72) Inventor Ginko Shimizu 3-3-3 Shinmachi, Ome-shi, Tokyo Inside of Toshiba Software Engineering Co., Ltd. (72) Inventor Hirofumi Goto 5-3-3-3 Shinmachi, Ome-shi, Tokyo Toshiba Software Engineering Co., Ltd. F-term (reference) 5B009 KA05 MA03 VB12

Claims (6)

[Claims] 1. A Chinese input conversion processing device for inputting Pinyin, which is a Chinese reading notation, and converting the input Pinyin into Chinese / Kanji characters, wherein the character string to be converted is input by Pinyin. Means for determining a range that needs to be corrected in the character string input by the input means, and correcting the character string in the range to create a valid correction character string candidate as Pinyin; and A conversion means for converting a candidate for a corrected character string obtained by the input correction means into Chinese characters, and an output means for outputting the conversion candidates obtained by the conversion means in a predetermined order. Chinese input conversion processor. 2. The apparatus according to claim 1, further comprising: a similar pronunciation detecting means for detecting, from among the corrected character string candidates obtained by the input correcting means, a pronunciation similar to the character string before correction. 2. The Chinese input conversion processing device according to claim 1, wherein conversion candidates obtained from the corrected character string detected by the detection means are output with priority. 3. An apparatus according to claim 2, further comprising: a correction learning detecting means for detecting a correction character string candidate obtained by said input correcting means, wherein the same input correction has been performed in the past. 2. The Chinese input conversion processing device according to claim 1, wherein conversion candidates obtained from the corrected character string detected by the means are output with priority. 4. A typo detecting means for detecting, from among candidates of a corrected character string obtained by said input correcting means, a character string input by a key close to a key corresponding to a character string before correction, 2. The Chinese input conversion processing device according to claim 1, wherein the output means outputs the conversion candidates obtained from the corrected character string detected by the typo detection means with priority. 5. A Chinese input conversion method for inputting Pinyin, which is a Chinese reading notation, and converting the input Pinyin into Chinese kanji, wherein a character string to be converted is input in Pinyin. At that time, determine the range of the input character string that needs to be corrected, correct the character string within that range, create a candidate for a corrected character string that is valid as Pinyin, and select these corrected character string candidates. A Chinese input conversion processing method, which performs conversion processing into Chinese kanji and outputs conversion candidates obtained by the conversion processing in a predetermined order. 6. When a character string to be converted is input to a computer having a function of inputting Pinyin, which is a Chinese reading notation, and converting the input Pinyin into Chinese kanji, A process of determining a range in the input character string that requires correction, a process of correcting a character string within the range to create a candidate for a corrected character string that is valid as Pinyin, and a candidate of these corrected character strings. A computer-readable recording medium which stores a program for executing a process of converting a character into a Chinese character and a process of outputting a conversion candidate obtained by the conversion process in a predetermined order.