JPS59220827A - Information processor with erroneous input correcting function - Google Patents

Abstract

PURPOSE:To perform correction processing with high efficiency without increasing its storage capacity by sectioning an input character string into (n)-character blocks when the input character string is not stored in a word dictionary, and referring to frequency information and substituting a character with a high evaluation value as a correct character. CONSTITUTION:A character string signal A inputted from a KANA (Japanese syllabary) keyboard 1 is passed through a character string converter 18 which converts special character concatenation into one special character to supply a signal B to a CPU12. A character string restoring device 19 restores a signal C to its original special KANA character string when the signal has a special character, and passes other signals as they are to a controller 4 as a signal D, which is sent as a signal E to the word dictionary to be looked up. A signal F when the word is stored in the dictionary 4 is sent to an output storage circuit as a signal G through a controller, and a signal N when not is outputted. The CPU12 when receiving the signal N sections the temporarily stored signal B into (n)-character blocks and sends them as a signal J to a frequency table 17 to obtain a frequency signal K; when a special character is contained, a special character freuqncy table 20 is looked up. The CPU12 calculates an evaluation value and substitutes a character with a maximum value as a correct character.

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to an information processing device, and in particular, to an information processing device such as a computer or a word processor, the present invention relates to an information processing device such as an electronic computer or a word processor. The present invention relates to an information processing device having a function of correcting errors.

2. Description of the Related Art Conventionally, the apparatus shown in FIG. 1 has been widely used as an information processing apparatus for inputting and processing sentences, for example, Japanese sentences, using words as units. Here, l is an input device, such as a kana keyboard, through which the operator inputs text. 2 is a central processing unit (C) that processes the input text and controls each part.
:PU), 3 is a word dictionary developed in the storage area of the external storage device, 4 is a controller that controls the external storage device under instructions from the CPU 2, and 5 is an output device that outputs and stores processed sentences. , 6 is an alarm display device that warns the operator of erroneous input.

The middle input word input from the Kana keyboard L is C.
The PU 2 determines via the controller 4 whether the word is stored in the word dictionary 3 or not. The input word is word dictionary 3
If the word is stored in the word, the CPU 2 determines that the word has been input correctly and sends the word to the output storage device 5.

On the other hand, if the input word is not stored in the word dictionary, the CPU 2 assumes that there is an error in the input, and sends an alarm signal to the alarm display device 6 as necessary. A warning indicator code for identifying that the word is not stored in the word dictionary 3 is added to the word and sent to the output storage device 5.

The entire desired sentence is processed by repeating the above-mentioned processing for the words that are sequentially input, and sending punctuation marks and the like from the ephemeral keyboard 1 to the output storage device 5 via the CPU 2.

However, in such conventional devices, a warning is displayed in response to an incorrect input, and a code is only attached to the incorrect input to identify it, which reduces the efficiency of text processing. It has the disadvantage that it decreases.

On the other hand, there is an information processing apparatus that has a function (erroneous input correction function) of correcting and outputting an incorrectly input word. FIG. 2 shows a conventional information processing device having an error input correction function, where 7 is a frequency table developed in a storage unit 1, for example, a ROM in the device, and n
Frequency information indicating how many times a combination of consecutive n characters (concatenation of n characters) appears in all words stored in the word dictionary 3 is stored for all concatenations of n characters, where is a natural number.

The device shown in the second figure performs the following correction process for the word when the word inputted using the kana keyboard 1 is not stored in the word dictionary 3 and the CPU 2 determines that the input is erroneous.

First, the CPU 2 sequentially divides the word determined to be incorrect into every n characters, from the first character to the nth character, from the second character to the n+1st character, and then concatenates the separated characters. are sequentially sent to the frequency table 7. Frequency table 7 shows the corresponding frequency information for each letter concatenation.
Supply to PU2. The CPU 2 evaluates each character in the word based on the frequency information, and determines, for example, the character with the lowest evaluation value as the character that was input incorrectly. Therefore, C
The PU 2 temporarily replaces the incorrectly input character with another character and refers to the frequency table 7 to obtain an evaluation value for that character. The CPU 2 performs a process of replacing the incorrectly inputted character with all other characters to obtain this evaluation value, determines the character with the highest evaluation value to be the correct character, and replaces the corrected character with the incorrectly inputted character. .

Next, the CPU 2 determines whether or not the word created as a result of the replacement is stored in the word dictionary 3.
If the dictionary 3 stores the word, it sends the word Jh as a correction result to the appearance storage device R5. If the word is not stored in word dictionary 3, CPIJ2
refers to the word dictionary 3, replacing the erroneously input characters in order of the character with the next highest evaluation value. If not all characters are stored in the word dictionary 3, a warning signal is sent to the warning display device 6, and the input word is sent to the word storage device 5 with a warning sign code attached. do.

In this way, the information processing device shown in the second diagram detects and corrects errors in input characters while referring to the frequency table for concatenations of n characters, and calculates a natural number n indicating the number of concatenations of characters. By increasing the correction capacity of the information processing device, it is possible to improve the correction ability of the information processing device.

However, the number of possible combinations of characters is m+1, where 1m is a natural number indicating the number of types of characters, and the storage capacity required for the frequency table 7 increases in proportion to C. Therefore, if n is increased, the problem arises that the device must be made larger.

Therefore, if the frequency table 7 is stored in an external storage device similar to the word dictionary 3, a problem arises in that external access takes time and processing time increases.

Generally, many of the information processing apparatuses shown in FIG. 2 having such an error correction function have been developed with the intention of manually writing sentences in English or the like. For example, in the case of English, the alphabet has 26 letters, so m = 26, and even if we set n = 3 and use a frequency table for concatenation of 3 letters, the required memory capacity is proportional to 263 = 17.57El. good.

In contrast, in the case of Japanese, there are 58 types of gana characters, including voiced and handakuten, m = 58. Since the storage capacity of the frequency table for n=3 is required to be proportional to 583-1!95,112, an external storage device with a significantly larger capacity is required. As mentioned above, in order to shorten the processing time, it is preferable to provide the frequency table 7 as the main memory of the CPU 2 instead of using an external storage device, but even in terms of the required storage capacity, it is This is extremely difficult, especially impossible in the case of information processing devices such as small computers.

In addition, in the information processing equipment that serves Japan 814,
Conventionally, for example, the N-B point and the hand-dakuten are input and processed as one independent character, so each of the voiced and half-voiced sounds is a concatenation of two characters. Therefore, the correction processing for a concatenation of n characters using a frequency table will essentially lose its ability to correct a concatenation of (n-1) characters if one voiced or handakuten is included in the n characters. Become. In this case, in order for the information processing device to have the same correction processing capacity as the concatenation of n characters, the storage capacity of the frequency table must be proportional to m''+1, and the required storage capacity is even greater. Furthermore, when processing a word with two consecutive voiced sounds, such as a φ word such as ゛gagagu (gagaku), it takes at least ゛? + 11 fl to correctly judge this.
> We need a frequency table for which four-letter concatenation. That is, in this case, the storage capacity is 584=11.316,4
98, and an exponentially larger storage device is required, so even if the frequency table is stored in an external storage device,
There was a problem in that the device became extremely large.

The present invention was made in view of such conventional problems, and
Its purpose is to create character combinations such as voiced and semi-voiced sounds that can be considered as one character, as well as character combinations consisting of specific character sets that frequently appear in Japanese sentences (hereinafter referred to as special character combinations).
is a special character that can be processed as a single character in the same way as the l^ sound other than the 'FiJ sound, etc., and by providing a special character concatenation table of n characters including these special characters, there is no need to increase the storage capacity. In order to achieve the above object, the present invention provides an information processing device having an error input correction function that can perform error input correction processing with high efficiency. If a permutation exists, the input word is converted to replace a specific permutation of characters with a specific code of one character, and a character string corresponding to the input mon, fro is generated. a string generation means, a word storage means for storing a plurality of words, and an il frequency information storage means for storing the appearance frequency in the carved word storage means for concatenation of characters consisting of a predetermined number of characters not including a specific code. , a second frequency information storage means for storing the frequency of appearance in the 70 words of condolence for a concatenation of characters consisting of a predetermined number of characters including at least one specific code; a reference means for referring to the 11-word storage means for the column; and, if the retrieved character string does not exist in the single crystal storage means as a result of the reference, a character consisting of a predetermined number of characters in order from the beginning of the retrieved character string; a decomposition means for decomposing into concatenations of each different character, a frequency information determining means for obtaining each different frequency information by referring to the first or second frequency information storage means for each concatenation of different characters; From the frequency information for the concatenation, calculate the evaluation value for each character that makes up the imported character string, extract the incorrectly entered characters based on the evaluation value, and replace the incorrectly entered characters with the correct character. The present invention is characterized by comprising a correction means for correcting the imported character string.

Hereinafter, the present invention will be described in detail with reference to the drawings.

FIG. 3 shows an example of the configuration of an information processing device having the erroneous input correction function of the present invention. Here, corresponding parts are given the same reference numerals and explanations are omitted for those that can be configured in the same way as before. do. In the device shown in FIG. 3, a character string converter 18 is provided between the kana keyboard 1 and the CPU 12 to convert special character concatenations into one special character, and a character string converter 18 is used to convert a special character concatenation into one special character between the CPU 12 and the controller 3. A frequency table 20 for special characters is arranged in parallel with a character string reducer 18 for reducing special character concatenations and a frequency table 17 for the clear sounds of n characters similar to the frequency table 7 described above. In addition to controlling each part, the CPU 12
The processing procedure shown in FIG. 4 is stored.

The input character string signal A output by the kana keyboard 1 is a character.

Column converter 18. The character string converter 18 converts a specific kana character string pre-contained in the input character signal A, that is, a concatenation of special characters, into unused special characters other than kana characters. do. For example, when the sulone 1 sha 9 exists in the character string signal A, if it is specified that the ゛sha' is to be converted to ISl, the character string converter The character string signal B outputted by 18 becomes 9S ko'. Further, for example, if 1 G' exists in the character string signal A, it is determined that it is converted to (G').The character string signal B is led to the CPU 12 and is temporarily stored. The resulting character string signal B is supplied to the character string reducer 18 as a character string signal C for each middle fill.

If there is a special character in the character string signal C, the character string reducer 19 reduces it to a specific kana character string, passes other characters as is, and sends the character string No. 46 to the controller 4. Output. That is, in the example described above, the character string reducer 19 reduces 'Sko' to 'Shakoa'. Note that if the application dictionary 3 includes special characters, the specific kana character string and the contents of the dictionary can be directly compared, so in that case, the character string reducer 19 may not be provided.

The character string signal is ilu + Goki 4 (Koyori, rlj J
It is converted into the character string No. 311 E that refers to the memory of li diction - profanity.

iJi, Rj Dictionary 3 outputs the character string signal E as it is as a dictionary output signal F if the manually inputted character string signal E and the rotated character string are stored as a Φ word, and if not stored, outputs the manually inputted character string signal E as a dictionary output signal F. A signal N indicating that the character string signal E is not registered in the dictionary is output. Dictionary output signal F
is converted by the controller 3 into a dictionary output signal G that conforms to CPt112, and when the CPU 12 receives the dictionary output signal G, it supplies it as an output character string signal H to the output storage device. When the CPU 12 receives the signal H, the CPU 12 supplies the alarm signal I to the alarm display device 6 as necessary to display an alarm, and performs the following error input correction process.

FIG. 4 shows an example of the procedure for correcting erroneous input.

First, in step St, when the CPt 112 receives the character string signal B, it temporarily stores the date of the signal, and
In step S2, the word dictionary 3 is referred to via the character string reducer 18 and the controller 4 to determine the input word. When the dictionary output signal G is sent from the word dictionary 3 via the controller 4, an affirmative determination is made in step S3, the word is output to the output storage device 5, and the process ends.

If a negative determination is made in step S3, that is,
When the CPU 12 receives the signal N, the CPU 12 divides the temporarily stored character string signal B including the erroneously input characters into every n characters and sequentially outputs them as an n character signal J. For example, n=2
As, the string signal B is! If it is a 1-ketekeno in which ``of'' of ``keiken'' is incorrectly input, 1-kechi'' and ``tegedokeno'' are output as an n-character signal J. When the frequency table 17 receives the n-character signal J, it displays frequency information indicating how many times the character string corresponding to the n-character signal J appears in all the words stored in the word dictionary 3. Output sequentially as . If the above-mentioned special character is included in the n-character signal J, the CPU 12 refers to the special character frequency table 20, and in this case, the frequency signal for the n-character signal J includes CPIJ12 $ from the special character frequency table 20. This is supplied.

The GPUI2 sequentially receives the M degree signal K, and sends the step S4.
calculates the evaluation value of the correct possibility for each character in the word from those values, and calculates the evaluation value of the correct possibility for each character (
(The character with i is determined to be an incorrectly input character. Therefore, in step S5, that character is replaced with one other character, and an evaluation value for the replaced character is calculated.CP
U12 obtains this evaluation value 114 for all kana characters other than the erroneously input character, and in step S8 tentatively determines that the character with the largest evaluation value is a correct character.

Next, in step S7, the CPU 12 corrects the incorrectly input characters by temporarily replacing them with characters determined to be correct, and outputs the corrected word as a new character string signal C.
In step S8, it is determined whether the word is stored in the word dictionary 3 using the same method as in step S3. If an affirmative determination is made here, the character string signal H corresponding to the word is sent to the output storage device 5, and the correction process is completed. On the other hand, if it is not stored in the word dictionary 3, in step 510, the characters that are tentatively determined to be regular characters are excluded, and the process returns to step S6, where the character with the next highest evaluation value is temporarily set as a regular character. Repeat the same steps. By repeating the above process, the output storage device 5 stores a sentence in which input errors have been corrected.

The storage capacity required for Class 1 Tables 17 and 20 will be explained in detail. The storage capacity required for the device of the present invention is the total storage capacity of the frequency table 17 and the special character order IW table 20, and if there are different types of special characters, then
(m0k) Proportional to I+.

For example, consider processing a character string consisting of two IQ sounds. The Kana keyboard outputs the high-pitched sound of ゞga'na as a ☆ character + a dot, that is, 1 ka' + ``zXゝ, or according to the present invention, a character + a voiced mark is output as one special character. In Japanese, there are 20 types of voiced sounds, so the storage capacity required in this case is (
57+20)2=5929. On the other hand, in the conventional information processing device G shown in FIG.
Since they are treated as 4 characters (Dakuten Jumonji + Dakuten), a frequency table corresponding to the 4 characters is required, and in this case, the storage capacity is 584.
=11. ．． 318,498 is required. in this way,
According to the invention, erroneous input nailing processing can be performed with a significantly smaller storage capacity than conventional devices.

As explained above, according to the present invention, the correction process for incorrectly input character strings can be realized using the reporter/server device I''i, which has a much smaller capacity than conventional devices. Not only can the 4,11 structure be made small, but also correction processing can be performed at high speed.

In the above embodiment, the input/output devices are described as a kana keyboard and an output storage device, respectively, but the present invention does not depend on the form of the input/output devices. That is, it can be applied to various uses by changing the input/output equipment. For example, as an input device? If a key voice recognition device is used, the information processing device can be made to have a voice input error correction function, and if an information transmission device is used as an output device, it can also be made to be an information processing device having a human input error correction transmission function. Furthermore, it goes without saying that the present invention can be easily applied to natural pi-speak other than Japanese. Furthermore, it is of course possible to easily deal with the case where there is one or more character errors in the input word.

[Brief explanation of drawings]

FIG. 1 is a block diagram showing an example of the configuration of a conventional information processing device, FIG. 2 is a block diagram showing an example of the configuration of a conventional information processing device having an error input correction function, and FIG. 3 is a block diagram showing an example of the configuration of a conventional information processing device having an error input correction function. FIG. 4 is a block diagram showing an example of the configuration of an information processing apparatus having an erroneous input correction function, and a flowchart showing an example of the erroneous input correction processing procedure. l... Kana keyboard, 2... Central processing unit (CPU), 3... Word dictionary, 4... Controller, 5... Output storage device, 6... Alarm display device, 7. ... Frequency table, 12... Central processing unit (CPtl), 17... Frequency table, 18... Character string converter-1 13... Character string reducer. 20... Special character frequency table, A-F, N... Signal. Patent applicant: Japan Broadcasting Corporation

Claims (1)

[Scope of Claims] 1) When a specific permutation of characters exists in the input word, conversion of the human-generated word to replace the specific permutation of characters with a specific code of one character. go and
a character string generation means for generating a character string corresponding to the input word; a word storage means for storing a plurality of words; and the word storage for concatenation of characters consisting of a predetermined number of characters not including the specific code. a first frequency information storage means for storing an appearance frequency in the word storage means; and a first frequency information storage means for storing an appearance frequency in the word storage means for a concatenation of characters consisting of the predetermined number of characters including at least one specific gradient sign. 2, the frequency information storage means and the character string are taken in;
a reference means for referring to the word storage means for the imported character string; and if as a result of the reference, the imported character string does not exist in the word storage means, the reference means refers to the word storage means from the beginning of the imported character string; a decomposition means that sequentially decomposes the predetermined number of characters into character concatenations, and a frequency for obtaining each separate frequency information by referring to the first or second frequency information storage means for each different concatenation of characters; an information determining means, and an evaluation value for each of the characters constituting the captured character string is determined from the frequency information regarding the concatenation of each of the different characters, and based on the evaluation value, an incorrectly input character is extracted; An information processing apparatus having a manual error correction function, comprising: a correction means for correcting the input character string by replacing the incorrectly input characters with correct characters. 2. In the information processing device having the function of correcting human input errors as set forth in claim 1, the correction means includes a determination means for determining a character input by error based on the evaluation value, and a determination unit for determining a character input by error based on the evaluation value. an evaluation value calculation means that applies all other characters to the determined character and calculates an evaluation value for each of the other characters based on the kudzu frequency information obtained by the frequency information determination means; a replacement means for replacing the erroneously input characters in the order of the character with the highest evaluation value obtained by the evaluation value calculation means; An information processing apparatus having an error input correction function, comprising a character correction means for determining a replaced character as the correct character and correcting the input character string. 3) In the information processing device having an error input correction function as set forth in claim 1 or 2, the word is a Japanese word, and the specific character permutation is a voiced sound and a hanshu sound. An information processing device having an error input correction function, characterized in that the information processing device includes a specific character set that frequently appears in the Japanese text.