KR100191001B1 - High speed spelling modifying device - Google Patents

Abstract

The high speed calibration device and the method can allow the user to quickly and automatically correct the wrong input string in Korean to the correct string. To this end, the fast-calibration apparatus stores a word generation unit for generating an input word sequence having at least one word by inputting a character string designated by a document creator, and a table for calibration in which wrong words and correct words are symmetrically arranged. And a word search unit for searching the correct word sequence for the input word sequence from the word generation unit from the correction table, and all and part of the input word sequence from the word generation unit. And a word substitute to replace the correct word sequence from the word.

Description

High speed calibration device and method

1 is a flow chart illustrating a conventional high speed calibration method.

2 is a block diagram of a high speed calibration apparatus according to an embodiment of the present invention.

3 is a flow chart illustrating a fast calibration method according to an embodiment of the present invention.

4 is a block diagram of a computer system for performing the flowchart shown in FIG.

* Explanation of symbols for main parts of the drawings

10: word generator 12: memory

14: word search unit 16: word replacement unit

20: input device 22: auxiliary storage device

24: CPU 26: main memory

28: expansion slot 30: monitor

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a word-processor for creating a document, and more particularly, to a high-speed proofing apparatus and a method for allowing a document creator to correct a wrong input string at a high speed and automatically. It is about.

In Korean general documents, the wrong grammar is often misunderstood because of wrong grammar, or because of the pronunciation confusion of one more or less 'letter'. Input errors that occur during the input of the document are often overlooked by the indifference of the inputter. Until now, the only way to correct this type of error was to use a tool called a spell checker. The spell checker, however, does a so-called off-line process that scans the document at the end of the input and finds the wrong word. Although the spell checker has a way to check a word, it is also off-line. The spelling checker provided is at risk of creating a document with the wrong words by the user by forgetting it and not executing it after typing. As a complementary method, a Quick Correct method has been introduced, which uses an on-line method for quickly performing this error when a wrong word is made by an inputter. However, this tool is made in English and has many shortcomings.

Since the automatic calibration method is a tool made for English, it is not suitable for deadlocks such as Japanese and Turkish, including Korean. In other words, due to the nature of the deadlock words, the application of practically impossible in the Korean word that is attached to live action to create many variations. The term "deadlock" refers to a language that is a form of language, and that a grammatical relationship is expressed by its function by attaching an independence to a word or research.

The reason why the existing method is impossible to express the error in Korean can be seen by the processing procedure of the flowchart shown in FIG. In other words, here, the input word (S) is simply processed by comparing the entire buffer with the already created buffer. In this way, the words created by the combination of many investigations or endings that are written in Korean, the deadlock word, are used. There is no choice but to be helpless.

For example, let's take the case where we need to replace the word 'yes' in the Korean language with 'yes'.

In order to add to the existing method, there is an irrationality that you have to register for all the existing verbs such as 'went' and 'come' and 'come'. This way, if you have n verbs, you have to register n each. In addition, since the combination of word endings and auxiliary verbs such as 'zoning' and 'tense' is not taken into consideration, even if they are considered as a combination of these, it is not possible to deal with the huge memory waste in order to register the errors of such endings. . This is true in the same way not only in the error of the mother but also in the investigation.

Another big problem is that Korean is a combination of elementary, middle, and tertiary. For example, the user who inputs 'have' should be corrected to correct it as 'have', but there is no way to add it because the wrong part actually starts with the last letter of the letter. However, if you show all of these nouns for all kinds of verbs, you will run into memory limitations as before.

Finally, needless to say about spacing errors in Korean.

Table 1 shows the types of words that cannot be processed by the conventional method.

As such, the automatic calibration used in English has a problem that cannot be correctly corrected in Korean.

Accordingly, the present invention has been made to solve the above-mentioned conventional problems, and an object thereof is to provide a fast calibration apparatus and method capable of quickly and automatically correcting wrong input strings in Korean with correct strings.

In order to achieve the above object, a fast proofing device according to an embodiment of the present invention includes word generation means for generating an input word sequence having one or more words by inputting a character string designated by a document creator, and wrong words and correct words. A memory for storing symmetrically arranged correction tables, word search means for retrieving the correct word sequence for the input word sequence from the word generation means, and the input word sequence from the word generation means; Word replacement means for replacing all and a portion with the correct word sequence from the word search means.

The fast proofing method according to an embodiment of the present invention may include a first process of generating an input word sequence having one or more words by inputting a character string designated by a document creator, and correcting the correct word sequence for the input word sequence in a memory. And a second process of searching from a stored calibration table and a third process of replacing the input word sequence with a correct word sequence retrieved from the calibration table.

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

Quick correction is a more advanced way to input the wrong and correct words in the form of a table in an internal buffer or memory, and then use this buffer whenever the user enters a word. The form to be processed by referring to

The calibration table, which should be made for quick calibration, takes the structure of a string storage having a part capable of storing a wrong word and a part capable of storing a wrong word as shown in Table 2 below. You will prepare an array of this structure and sort the wrong words by keys. Wrong words in the orthodontic table are represented by wrong words, and correct words by correct words.

We will use the new '-' and '' commands to handle the types mentioned in Table 1.

'-' Is a command that indicates where the wrong word to be processed appears in the error clause. When a word that is a processing object is divided into real and fictitious words of deadlock, it is shown in Table 3 below. At this time, if part (b) is wrong, part (a) does not have to be related to any due diligence, so part (a) is represented by '-'. That is, in case of 'up', it appears as '-up'. On the contrary, suppose the part is wrong. In this case, since part of (b) does not mean any fiction, this part (b) can be represented by '-', so it becomes like a bulletin board.

In the quick correction, the latter case is recognized as the same even if the '-' is omitted. This is because most of the words in the word of affairs are clearly stated as being wrong rather than if a word is completely wrong. And since part (B) is always meaningless without meaning, this latter case is expressed by the law that applies if part (A) is in input clause. Also, the '-' part doesn't have to be due diligence or vanity, as it always needs to be filled with anything. In other words, it only needs to be filled with some letters.

Now explain the command. Quick corrections are supposed to recognize the incorrect words as being spaced. This is in line with the idea of recognizing two or more words as one wrong word. In this case, use this command to indicate the part of the word that is incorrect. For example, if 'down' and 'load' should always be added as 'download', the wrong word is 'download' and the correct word is 'download'.

As mentioned above, if two commands of '-' and '' are used, wrong words such as those shown in Table 1 can be registered as shown in Table 4 below. It is possible.

What is unique in Table 4 is a method of adding up to unchanged syllables in order to express (reject, refrain). As a result, the words 'reluctant, refrained, forgotten' etc. are corrected.

In this way, however, some input phrases may not be corrected by incorrect words registered with the '-' command (as in -b, -b). The processing is correct because it was corrected by 'stepping on' and the actual word needed to be corrected.However, the input word 'straight' is different from the registered wrong word and the stemming is different. to be. For these words, quick corrections have a rule to avoid corrections, and if you register them (as opposed to), they will be ignored without the correction of the wrong word '-b'.

In other words, if the wrong words and the correct words are the same as those found in the correction table, this law of avoidance applies.

The actual quick correction then processes the entered word while looking at the correction table created using these two commands. If the word entered is 'wrong', the first thing to look for in the calibration table is 'wrong'. If found, replace the wrong part with the correct word. However, in this example, because (a) is not wrong, we find (b) that it is wrong. The system looks for the entered word. Therefore, replacing this (b) part completes the process completely.

Spacing errors entered using the '' command are processed by checking the input word in the form of correction table in the form including '' first, and searching the input word after '' if it is not found. This is summarized in detail as follows.

Input clause S that can have a '' command

In this case, S is searched for in the calibration table, and the word is searched until S ends in one word while being further divided into words divided into ''. If you find it in the calibration table on the way, you can replace it with the correct words. If the S created after the input word is reversed can be processed in the same way. Now take this example. From a beautiful peninsula and a junior school, you can apply from the information presented in (Table 1). The first example finds a beautiful peninsula, a peninsula and a peninsula in the calibration table. At this time, since the second test (Korean Peninsula, Korean Peninsula) can be found, the calibration is successful. The second example is found in reverse order of {ㅓ ㅌ ㅜ ㅂ f ㅗ ㄹ ㅛ ㄱ ㄱ ㅎ ㄴ ㅁ ㄱ ㅜ ㄱ, f ㅓ ㅌ TT ㅜ ㅗ f, ㅓ ㅛ ㄱ ㅜ ㅂ,. Again, the second test (from-, from-) finds the search successful, so the calibration is complete. The inverse word is also the same as the S-shape expressed above, so the search method is ultimately the same.

However, it is necessary to limit the word that can be represented by '' to some extent, and the system decides the maximum number of words that can be generated as ''. Usually 1 is enough.

Referring to FIG. 2, a fast calibration according to an embodiment of the present invention includes a word generation unit 10 for inputting words for inputting words to be documented by a user, and a memory 12 storing a calibration table. The apparatus is described.

The word generator 10 plays a role of generating a word sequence of a type required by the system. To this end, the word generation unit 10 includes a keyboard for inputting a character string designated by a user, a word buffer sufficient to store a certain number of words, and an arithmetic unit having some calculation functions. . The word generation unit 10 sequentially stores the strings input by the user in a word buffer in itself, and checks whether the character input by the user is '' or a newline character (ie, a space bar or the enter key). . When the '' character is input, it is compared whether the number of allowable input phrases required by the system is equal to the number of inputs of the blank key. When the number of allowed input phrases and the number of blank key inputs are the same and when the newline character is input, the word generation unit 10 traces the contents accumulated in the input word buffer in reverse and inputs the input word (s). To complete. At this time, if the character looked after the newline is a newline character, the input word is determined from the position after the newline character to the position where the character starts to be traced backward. If the character is '', then the input sequence is generated similarly when the number of '' accumulated so far exceeds the number determined by the system.

The memory 12 includes a calibration table in which wrong words and corresponding words are arranged side by side in the form of Table 4 so that a user can correct a wrong input string (word). The calibration table is preset by the manufacturer.

The fast calibration device further includes a word search unit 4 and a word substitution unit 16 for commonly inputting the input word s from the word generation unit 10.

The word search unit 14 checks whether the strings of the input word sequence s are in the calibration table stored in the memory 12, and if there is a correct word in the found items (s-, r-). Modify the temporary input word buffer s with 'r-'. However, if the search fails, it generates a reverse word clause (sr) of the input word sequence again and searches for it because the first string becomes the end of the actual input word. Then, if (sr-, rr-) is found, it modifies the temporary input word buffer (s) with rr, which is also expressed as a correct word. If the character string is modified in the above process, it is immediately supplied to the word substitution unit 16.

Otherwise, the word search unit 14 repeats the search by excluding the previous word one by one in the input word sequence s and finds replaceable words to modify the temporary input word buffer s. . If there is a modified word or words in the above process, the words are supplied to the word replacement unit 16 in the modified temporary input word buffer s.

Lastly, when the corrected words are supplied from the word search unit 14, the word replacement unit 16 converts the input word sequence s from the word generation unit 10 to the word search unit 14. Generate corrected word sequences in whole or in part by replacing modified words from The word substitution unit 16 transmits the corrected word sequence via the output line 11.

3 is a flowchart of a fast calibration method according to an embodiment of the present invention, which is performed by the computer system shown in FIG. Before describing FIG. 3, the computer system shown in FIG. 4 will be described.

In FIG. 4, the computer system includes an input device 20, a main memory 26, and an auxiliary memory 22 which are commonly connected to a central processing unit 24 by a bus line. ) And an expansion slot 28.

The input device 20 inputs character strings and various commands designated by the document creator and transmits them to the CPU 24. The input device 20 may be a keyboard, a key matrix, a key switch device, or the like.

The auxiliary memory device 22 stores a word processor program and a calibration table as shown in Table 4 as well as the flowchart of FIG. 3 to be performed by the CPU 24. FIG. The main memory 26 separately stores documents and various data created by the CPU 24.

The main memory 26 temporarily stores documents and various data processed by the CPU 24 and the main memory 26 stores various programs (eg, words) to be executed by the CPU 24. The processor program, the flowchart of FIG. 3) and the calibration table as shown in Table 4 above are temporarily stored.

The expansion slot 28 functions to transfer image data and various data processed by the CPU 24 to peripheral devices connected to the outside, and to transfer information from peripheral devices to the CPU 24. .

The CPU 24 processes the character strings and various data from the input device 20 and temporarily stores them in the main memory 26 and communicates with the peripheral device by a command from the input device 20. Control of the auxiliary storage device 22 and the main memory 26. As an example of the control operation, the CPU 24 causes the program and various data stored in the auxiliary storage device 22 to be stored in the main memory 26, or the various data stored in the main memory 26 is stored. To be stored in the auxiliary storage device 22. In addition, the CPU 24 includes not only buffers and registers that can be used to perform the flowchart shown in FIG. 3, but also various buffers, mode flags, and registers sufficient to execute various programs.

The monitor 30 connected to the expansion slot 28 performs a function of displaying image data from the CPU 24 input via the expansion slot 28.

Next, the flowchart shown in FIG. 3 will be described in conjunction with the computer system shown in FIG.

In the thirty-second step, the CPU 24 sets an allowable blank value that designates the number of words that can be input to the allowable blank counter SC allocated among the registers in itself. The allowable blank value may be changed by a user's designation. After performing step 32, the CPU 24 inputs a character designated by the user from the input device 20 (step 34). The CPU 24 checks whether the input character is a key (Key, space bar or enter key) indicating completion of a word. When the input character is the space bar key, the CPU 24 adds the value of the blank key counter by one when the value of the blank key counter is small (step 36). The CPU 24 stores the input character in the position of the input word buffer C corresponding to the counter value of the character string counter Ci (step 38). The CPU 24 then adds the value of the character string counter Ci by one and returns to step 34. (Step 40).

On the other hand, when the input character is the enter key in step 36 or the value of the blank key counter is equal to the allowed blank value SC, the CPU 24 sets the value of the string counter Ci to 1. And subtracts the value of the subtracted string counter Ci into the word start position counter P (step 42). Then, it is checked whether the value of the word start position counter P is greater than or equal to zero (step 44). In step 44, when the value of the word start position counter P is greater than or equal to 0, the CPU 24 is located at the position of the input word buffer C corresponding to the word start position counter P. FIG. Check for missing characters (step 46). If there is a character in step 46, the CPU 24 checks whether the character is a newline character (i.e., enter key) (step 48). If it is not a newline character in step 48, the CPU 24 subtracts the value of the word start position counter P by 1 and returns to step 44 again (step 50). In step 46, if there are no characters, the CPU 24 compares whether the value of the allowed blank counter SC is greater than zero (step 52). In step 52, when the value of the allowed blank counter SC is greater than 0, the CPU 24 subtracts the value of the allowed blank counter SC by 1 (step 54). The CPU 24 subtracts the value of the word start position counter P by one (step 56). On the contrary, when the value of the word start position counter P is less than 0 in the 44th step, the value of the allowable blank counter SC is less than 0, or is a newline character in the 48th step, the CPU 24 adds the value of the word start position counter P by one (step 58).

The CPU 24 calculates the word length by subtracting the value of the word start position counter P from the value of the character string counter Ci and records the calculated word length in the word length counter len. (Step 60). In addition, the CPU 24 checks whether the value of the word length counter len is greater than zero (step 62). After performing step 62, the CPU 24 initializes the value of the check character counter j to 0 (step 64). Thereafter, the CPU 24 increments the value of the word start position counter P and the value of the check character counter j by 1 (steps 68 and 70), and the check character counter j. (Step 72) In step 72, if the value of the check character counter j is smaller than the word length len, the CPU determines whether or not the value of the word length counter len is smaller than the word length len. Step 24 repeats step 72 from step 66 to sequentially store check target words on the input word buffer C in the check word buffer S.

If the value of the check character counter j is not smaller than the word length len in step 72, the CPU 24 initializes the value of the check start position counter k to zero. step). The CPU 24 searches the calibration table stored in the memory 26 so that words on the check word buffer S after the position corresponding to the value of the check start position counter k are added to the check table. Check for presence (step 76). When the words are not present in the calibration table in step 76, the CPU 24 generates the inverse words s-, * of the words (step 78). The CPU 24 searches the calibration table and checks whether there is a reverse word (s-, *) in the calibration table (step 80). When the reverse word (s-, *) is not found in step 80, the CPU 24 compares the value of the test start position counter k with the word length len to search further words (Step 82). When there is a word to be searched in step 82, the CPU 24 selects the remaining words except for the first word among the currently checked words in the check word buffer S as the target word S '. Set (step 84), add the number of characters of the excluded word and 1 to the test start position value len to calculate a new test start value, and calculate the new test start value into the test start position counter ( k) (step 86). After execution of the 86th step, the CPU 24 returns to the 76th step.

On the other hand, when the words (s-, *) are found in the calibration table in step 76, the CPU 24 performs a correct word sequence r- on the calibration table for the words s-. (Step 88). When the reverse word Sr- is found in the calibration table in step 80, the CPU 24 selects the correct reverse word Rr- on the correction table for the reverse word Sr-. Read it (step 90). After performing the 88th and 90th steps, the CPU 24 retrieves the words on the input word buffer C after the position corresponding to the value of the check position counter k. Or replace with the searched correct word Rr- (step 92). The CPU 24 then records all or part of the replaced words in a document on the main memory 26 and returns to the thirty-second step.

On the other hand, if there are no words to search in step 82, the CPU 24 writes the words in the input word buffer C in a document on the main memory 26, and then returns to step 32. In addition, when the value of the word length counter len is less than zero in step 62, the CPU 24 returns to step 32.

As described above, the present invention provides an advantage that a document creator can quickly correct a document by automatically correcting a plurality of words by searching while inputting a character string of the document.

Claims (6)

A word generation unit for generating an input word sequence having one or more words by inputting a character string designated by a document creator, a memory for storing a table for calibration in which wrong words and correct words are symmetrically arranged, and the word generation means Word search means for retrieving the correct word sequence for the input word sequence from the correction table; and word substitution for replacing all and part of the input word sequence from the word generation means with the correct word sequence from the word search means; A high speed calibration device comprising means. The apparatus of claim 1, wherein the word generating unit generates the previously inputted character strings as the input word strings when the character input by the user is an enter key. The method of claim 1, wherein the word generating means compares the number of inputs of the blank key and the value of the allowable blank key when the character input by the user is a blank key, and as a result, the number of blank key inputs and the value of the allowed blank key are the same. The case of the high-speed calibration device, characterized in that for generating previously input strings as an input word sequence. 4. The word phrase generating means according to claim 3, wherein the word generating means compares a keyboard for relaying with a user, an input word buffer for storing character data from the keyboard, and a number of inputs of the blank key and a value of the allowed blank key. A high speed calibration device comprising a device. 5. The fast calibration apparatus of claim 4, wherein the value of the allowable blank key is changeable by the user to any value. A first step of generating an input word sequence having one or more words by inputting a character string designated by a document creator; a second step of retrieving a correct word sequence for the input word sequence from a calibration table stored in a memory; And a third step of replacing word sequences with correct word sequences retrieved from the correction table.