JPH11259465A - Character string prediction device and method therefor - Google Patents

Abstract

PROBLEM TO BE SOLVED: To obtain the candidate of a predicted character string and to efficiently input the character string even if a user does not clearly indicate the start of input prediction by permitting the user to predict input when the recognition result of a character pattern cannot be recognized. SOLUTION: An inputted character pattern is recognized and recognized candidate characters obtained as recognition results are sequentially stored in an input buffer (steps 201-204a). When the input character pattern cannot be recognized, a prediction operation is started and a character string having the character string for a head part is retrieved from a word dictionary based on the character string stored in the input buffer (steps 203-205). Thus, even if the user does not clearly indicate the start of input prediction, the candidate of the predicted character string is obtained and the character string can efficiently be inputted.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a character string predicting apparatus for predicting a character string to be subsequently inputted by a user based on a character string having a pattern input, and more particularly to a display timing of prediction candidates. The present invention relates to a character string prediction device and a character string prediction method having features.

[0002]

2. Description of the Related Art Input prediction (character string prediction) is defined as the time when the first few characters are input, as reported in the literature (the 48th Information Processing Society of Japan National Convention Proceedings 4J-11 and others). This is a character input method of searching a word dictionary or the like and extracting a character string (word) starting with a few characters.

For example, when a character string “kanji dictionary” is input, when the user inputs “kanji”, the “kanji”
Is predicted as a character string subsequent to, and finally a "Kanji dictionary" is output. Generally, this input prediction is performed by the user explicitly instructing prediction, such as pressing a “prediction button” or drawing a “horizontal line” following an input character string.

[0004]

The purpose of input prediction is to allow a user to select a subsequent character string predicted from an input character string, thereby complementing a character string intended for input and reducing input operations. It is in. However, explicitly instructing the start of input prediction increases the number of operations required for input, contrary to the original purpose.

[0005] It is also conceivable to start input prediction automatically when an arbitrary number of characters are input.
Considering the processing time required for character recognition and input prediction, inputting characters one after another while waiting for the result of character recognition and input prediction can shorten the input time, and may hinder the user's character input operation. There is.

SUMMARY OF THE INVENTION The present invention has been made to solve the above-mentioned problem, and performs input prediction when a result of recognition of a character pattern input by a user becomes unrecognizable.
Provided is a character string prediction device and a character string prediction method that can obtain a candidate for a predicted character string and efficiently input a character string without a user explicitly instructing the start of input prediction. Aim.

[0007]

A character string predicting apparatus according to the present invention is a character string predicting apparatus for predicting a subsequent character string based on an input character string. A recognition dictionary that stores information necessary for pattern recognition, and a recognition unit that performs a recognition process on a character pattern input by the input unit with reference to the recognition dictionary and outputs recognition candidate characters corresponding to the input character pattern. When,
An input buffer for sequentially storing recognition candidate characters output from the recognition means as input characters, a word dictionary storing information of each word, and an input character pattern processed as an unrecognizable character by the recognition means. Prediction control means for searching the word dictionary for a character string having the character string at the head thereof based on the character string stored in the input buffer; and a character string obtained by the prediction control means as a candidate for a predicted character string. And candidate display means for displaying as

According to this configuration, when the input character pattern becomes unrecognizable, a prediction operation is started based on the character string input so far, and a candidate for the predicted character string is displayed. You. Therefore, if the input character pattern can be recognized, character input is continued as it is,
When recognition becomes impossible, a candidate for a predicted character string can be obtained to complement a character. Thus, even if the user does not explicitly instruct the start of the input prediction, the input of the character string can be efficiently performed while obtaining the candidate of the predicted character string as needed.

When the input character pattern is processed as an unrecognizable character by the recognition means, the prediction control means determines whether or not the character string stored in the input buffer satisfies a predetermined condition. When a predetermined condition is satisfied, a character string having the character string at the head is searched from the word dictionary.

According to such a configuration, when the input character pattern becomes unrecognizable, if the character string input up to that time satisfies a predetermined condition (a predetermined number of characters), The prediction operation is started based on the input character string. Therefore, when the recognition becomes impossible, it is possible to solve a problem that the number of input characters at that point is small and the prediction candidates cannot be narrowed down and a large number of prediction candidates are displayed. As a result, the user can obtain a candidate for the predicted character string narrowed down by a certain presentation and complement the character.

A selecting means for selecting a candidate of a predicted character string displayed by the candidate display means; a character corresponding to the unrecognizable character is specified from the predicted character string selected by the selecting means; And a recognition learning means for associating the set character with the input character pattern processed as the unrecognizable character and registering it in the recognition dictionary.

According to such a configuration, a character corresponding to an unrecognizable character is obtained from the predicted character string selected by the user, and the dictionary registration for the character is performed. Therefore, the character that has become unrecognizable will be recognized from the next time, and the input efficiency can be further improved.

[0013]

Embodiments of the present invention will be described below with reference to the drawings. FIG. 1 is a block diagram showing the configuration of the character string prediction device of the present invention. Note that the present apparatus has a pattern recognition function and predicts a subsequent character string based on a character string input with a pattern. Hereinafter, handwritten character recognition will be described as an example of pattern recognition. Also,
The present apparatus is realized by a computer which reads a program recorded on a recording medium such as a magnetic disk and the operation of which is controlled by the program.

In FIG. 1, an input device 101 is for inputting a character string. Here, a handwriting tablet is used in which an input character pattern is obtained as coordinate values.

The character input unit 102 cuts out the input character pattern, performs a process of recognizing the character pattern in the pattern recognition unit 103, and stores the recognition candidate character obtained as a result of the recognition in the input buffer 106.

The pattern recognition section 103 recognizes an input character pattern by referring to the pattern recognition dictionary 104, and outputs recognition candidate characters as a result of the recognition. The third
In the case of the embodiment, when character recognition is performed, a pattern learning dictionary 105 is also referred to in addition to the pattern recognition dictionary 104. The pattern recognition dictionary 104 stores information necessary for pattern recognition, and is used as a standard dictionary. The pattern learning dictionary 105 stores information on character patterns actually input by the user, and is used together with the pattern recognition dictionary 104 as an extended dictionary during character recognition. The input buffer 106 sequentially stores the recognition candidate characters obtained by the pattern recognition unit 103 as input characters.

The prediction control unit 107 controls the entire character string prediction process. In the first embodiment, the prediction control unit 107 monitors a character pattern input by the character input unit 102,
When the input character pattern becomes unrecognizable by the pattern recognizing unit 103, the dictionary retrieving unit 109 is instructed to start a prediction operation. Further, in the second embodiment, the prediction control unit 107 monitors a character pattern input in the character input unit 102, and when the input character pattern becomes unrecognizable in the pattern recognition unit 103, the prediction determination unit 107 108, it is determined whether the character string input so far in the input buffer 106 satisfies the condition of the base character string (the number of characters set in advance). If so, an instruction is issued to the dictionary search unit 109 to start a prediction operation.

Further, the prediction control unit 107 controls the word dictionary 1
11 presents a candidate for the predicted character string retrieved from the output buffer 11, and outputs a candidate for the predicted character string selected by the user to the output buffer 1.
14 is stored.

In accordance with an instruction from the prediction control unit 107, the prediction determining unit 108 determines whether or not the character string stored in the input buffer 106 satisfies a preset base character string condition. This prediction determination unit 108 is used in the second embodiment. The condition of the base character string is the number of characters of the input character string.
If it is longer than characters, it will be the base character string.

In accordance with an instruction from the prediction control unit 107, the dictionary search unit 109 sets a character string stored in the input buffer 106 as a base character string, and outputs a predicted character string having the base character string as a head part from the word dictionary 111. A search is performed, and the candidate of the predicted character string is stored in the candidate buffer 112.

The recognition learning unit 110 specifies a character corresponding to an unrecognizable character from the predicted character string selected by the user in accordance with an instruction from the prediction control unit 107, and the code data of the character has become unrecognizable. The stroke data of the character pattern is registered in the pattern learning dictionary 105 in association with the stroke data. This recognition learning unit 110 is used in the third embodiment.

The word dictionary 111 stores information such as a headline character string and a read character string for each word. The candidate buffer 112 stores the predicted character string candidates obtained by the dictionary search unit 109.

The output control unit 113 includes an output buffer 114
Is displayed on the output device 115 (predicted character string candidates selected by the user). The output buffer 114 stores a predicted character string candidate selected by the user.

The output device 115 is, for example, a CRT (Cathod
It is a screen display device such as an e-ray tube) or an LCD (Liquid Crystal Display), which allows the user to interactively proceed with character input processing.

Next, the operation of the present apparatus will be described. First, a processing operation according to the first embodiment will be described. FIG. 2 is a flowchart showing a flow of processing from character input to output of a predicted character string in the first embodiment of the present invention.

The handwritten character pattern input from the input device 101 is cut out in the character input unit 102 (step 201). The cut-out character pattern is recognized by the pattern recognition unit 103 with reference to the pattern recognition dictionary 104 (step 201). If the input character pattern is recognized (No in step 203),
Recognition candidate characters obtained as a result of the recognition are stored in the input buffer 106 (step 204), and the recognition candidate characters are displayed in the editing area on the screen as input characters (step 204a).

Here, the prediction control unit 107 monitors pattern input and pattern recognition processing, and when the input character pattern becomes unrecognizable, instructs the dictionary search unit 109 to start a prediction operation.

That is, when the input character pattern is processed as an unrecognizable character (Yes in step 203).
s) In response to an instruction from the prediction control unit 107, the dictionary search unit 109 sets a character string having that character string at the beginning as a predicted character string, based on the character string currently stored in the input buffer 106, as the word dictionary 111. Search from (Step 2
05), the candidate of the predicted character string obtained by the dictionary search is stored in the candidate buffer 112 (step 20).
6).

The prediction control unit 107 includes a candidate buffer 112
Are displayed in the prediction candidate area on the screen, and the user performs candidate selection processing (step 207). When the predicted character string is selected by the user, the predicted character string is stored in the output buffer 114 (Step 208), sent to the output device 115 in the output control unit 113, and displayed on the screen (Step 20).
9). Finally, each buffer is initialized for the next input (step 210).

As described above, in the first embodiment, when an input character pattern becomes unrecognizable, a prediction operation is started based on a character string input so far. Therefore, if the input character pattern can be recognized, the character input can be continued as it is, and if the input character pattern cannot be recognized, a candidate for a predicted character string can be obtained to complement the character.
As a result, even if the user does not explicitly instruct the start of input prediction, a candidate for a predicted character string can be obtained as needed, and a character string can be input efficiently.

Next, a processing operation according to the second embodiment will be described. FIG. 3 is a flowchart showing a flow of processing from character input to output of a predicted character string in the second embodiment of the present invention. Steps 301 to 304a
Are the same as steps 201 to 204a in FIG.

That is, a pattern of a handwritten character input from the input device 101 is cut out by the character input unit 102 (step 301). The cut-out character pattern is recognized by the pattern recognition unit 103 with reference to the pattern recognition dictionary 104 (step 301). If the input character pattern is recognized (step 303
No), the recognition candidate character obtained as a result of the recognition is stored in the input buffer 106 (step 30).
4) The recognition candidate character is displayed in the editing area on the screen as an input character (step 304a).

Here, the prediction control unit 107 monitors the pattern input and pattern recognition processing, and when the input character pattern becomes unrecognizable, instructs the prediction determination unit 108 to store the current character pattern in the input buffer 106. It is determined whether or not the given character string satisfies the condition of the base character string (step 305). The condition of the base character string is the number of characters of the input character string. Here, a character string of two or more characters is set as the base character string.

As a result of the determination, when the condition of the base character string is satisfied (Yes in step 305), the prediction control unit 107 instructs the dictionary search unit 109 to start a prediction operation. As a result, the dictionary search unit 109 sets the input buffer 10
6, a character string having the character string at the head thereof as a predicted character string based on the character string currently stored in the word dictionary 1
11 (step 306), and stores a candidate for a predicted character string obtained by the dictionary search in the candidate buffer 112 (step 207).

On the other hand, when the condition of the base character string is not satisfied (No in step 305), the prediction control unit 107
Directly stores the character string stored in the input buffer 106 in the output buffer 113 (step 309).

The subsequent processing is the same as that of FIG. 2, and the prediction control unit 107 displays the candidate of the predicted character string stored in the candidate buffer 112 in the predicted candidate area on the screen, and the user selects the candidate. Is performed (step 308). When the predicted character string is selected by the user, the predicted character string is stored in the output buffer 114 (step 30).
9), sent to the output device 115 in the output control unit 113 and displayed on the screen (step 310). Finally, each buffer is initialized for the next input (step 31).
1).

As described above, in the second embodiment, when the input character pattern becomes unrecognizable, if the character string input up to that time satisfies the preset number of characters, the input Start a prediction operation based on a character string. Therefore, when the recognition becomes impossible, it is possible to solve a problem that the number of input characters at that point is small and the prediction candidates cannot be narrowed down and a large number of prediction candidates are displayed. As a result, the user can obtain a candidate for the predicted character string narrowed down by a certain presentation and complement the character.

Next, a processing operation according to a third embodiment will be described. FIG. 4 is a flowchart showing the flow of processing from character input to output of a predicted character string in the third embodiment of the present invention.

The pattern of the handwritten character input from the input device 101 is cut out in the character input unit 102 (step 401). The cut-out character pattern is recognized by the pattern recognition unit 103 with reference to two dictionaries, a pattern recognition dictionary 104 and a pattern learning dictionary 105 (step 402). If the input character pattern is recognized (No in step 403), the recognition candidate character obtained as a result of the recognition is stored in the input buffer 106 (step 404), and the recognition candidate character is displayed on the screen as an input character. (Step 404a).

Here, the prediction control unit 107 monitors pattern input and pattern recognition processing, and when the input character pattern becomes unrecognizable, instructs the dictionary search unit 109 to start a prediction operation.

That is, when the input character pattern is processed as an unrecognizable character (Yes in step 403)
s) In response to an instruction from the prediction control unit 107, the dictionary search unit 109 sets the character string having the character string at the beginning as the predicted character string based on the character string currently stored in the input buffer 106, and sets the word dictionary 111 as the predicted character string. Search from (Step 4
05), the candidate of the predicted character string obtained by the dictionary search is stored in the candidate buffer 112 (step 40).
6).

The prediction control unit 107 includes a candidate buffer 112
Are displayed in the prediction candidate area on the screen, and the user performs candidate selection processing (step 407).

Here, in the third embodiment, when a candidate of a predicted character string is selected by the user, the recognition learning unit 1
Reference numeral 10 identifies a character corresponding to an unrecognizable character from the predicted character string selected by the user, and associates the code data of the character with the stroke data of the unrecognizable character pattern and registers it in the pattern learning dictionary 105. (Step 408).

The candidate of the predicted character string selected by the user is stored in the output buffer 114 (step 40).
9) In the output control unit 113, the data is sent to the output device 115 and displayed on the screen (step 410). Finally, each buffer is initialized for the next input (step 41).
1).

As described above, in the third embodiment, when an input character pattern becomes unrecognizable, a prediction operation is started based on the input character string, and further, a predicted character selected by the user is selected. The character corresponding to the unrecognizable character is obtained from the column and registered in the dictionary. Therefore, in the event that recognition becomes impossible, a candidate for a predicted character string can be obtained to complement the character. In addition, by registering the dictionary, the character that has become unrecognizable will be recognized from the next time, and input efficiency will be reduced. It can be further improved.

Next, each of the above-described embodiments will be described using specific examples. FIG. 5 is a specific example for explaining input prediction in the first embodiment. In the drawing, reference numeral 1 denotes an input frame for inputting a handwritten character, 2 denotes a recognition result display unit for displaying a recognition result, and 3 denotes a prediction candidate display unit for displaying a candidate for a predicted character string. .

For example, as shown in FIG. 5 (a), when a character pattern input by handwriting has been normally recognized, the processing of character pattern input and pattern recognition is repeated, and I can't.

On the other hand, as shown in FIG. 3B, when the character pattern input by handwriting cannot be recognized, the prediction control unit 107 instructs the dictionary search unit 109 at this point to start the prediction operation. Thus, a predicted character string candidate is output. In this example, each character up to "con" was recognized, but the next character "pi" could not be recognized, and as a candidate for a predicted character string starting with "con" input up to that point,
"Concrete", "Contest", "Concorde",
"Concourse" ... "Computer" ... are displayed.

FIG. 6 is a specific example for explaining input prediction in the second embodiment. In the drawing, reference numeral 1 denotes an input frame for inputting a handwritten character, 2 denotes a recognition result display unit for displaying a recognition result, and 3 denotes a prediction candidate display unit for displaying a candidate for a predicted character string. . The condition as a base character string is the number of input characters, and a character string of two or more characters is a base character string.

In the second embodiment, when the input character pattern becomes unrecognizable, the prediction control unit 107 instructs the prediction judgment unit 108 to convert the character string stored in the input buffer 106 into the base character string. (2 characters or more) are satisfied. In the case of FIG. 6A, since the character input before the unrecognizable character is one character of "", the condition of the base character string is not satisfied. Therefore, no character string prediction is performed and no candidate is displayed.

On the other hand, in the case of FIG. 3B, the characters input before the unrecognizable character are two characters of "con", which satisfies the condition of the base character string. For this reason, the prediction control unit 107 instructs the dictionary search unit 109 to start a prediction operation and outputs a predicted character string candidate. In this example, two characters of "con" were recognized, but could not be recognized by the next character of "pi", and as a candidate of a predicted character string having "kon" at the beginning input up to that point, "concrete",
"Contest", "Concorde", "Concourse" ...
"Computer" ... is displayed.

FIG. 7 is a specific example for explaining input prediction in the third embodiment. In the drawing, reference numeral 1 denotes an input frame for inputting a handwritten character, 2 denotes a recognition result display unit for displaying a recognition result, and 3 denotes a prediction candidate display unit for displaying a candidate for a predicted character string. .

As shown in FIG. 7A, when the input character pattern becomes unrecognizable, the prediction control unit 107
Instructs the dictionary search unit 109 to start the prediction operation and outputs a predicted character string candidate. In this example, each character up to "con" was recognized, but the next character "pi" could not be recognized, and as a candidate for a predicted character string starting with "con" input up to that point, "Concrete", "Contest", "Concorde", "Concourse" ... "Computer" ... are displayed.

Here, as shown in FIG. 5B, when the user selects a candidate for a predicted character string, the recognition learning unit 110 causes the selected predicted character string candidate to correspond to an unrecognizable character. The character is specified, and the code data of the specified character and the stroke data of the character pattern that cannot be recognized are registered in the pattern learning dictionary 105. That is, if the user selects “computer” from the candidates for the predicted character string, the recognition becomes impossible at the third character, so “pi” of “computer” is to be registered. The character pattern input by the user is registered in the pattern learning dictionary 105 in association with the character pattern.

The character patterns registered in the pattern learning dictionary 105 will be referred to at the time of next recognition. That is, when the same character pattern is subsequently input, the input character pattern can be recognized with reference to the pattern learning dictionary 105.

The above embodiments have been described on the assumption that handwritten characters are input. However, the present invention is not limited to this, and can be applied to input methods generally based on a pattern recognition function, such as voice input.

In each of the above embodiments, the unrecognizable character is not only a character that could not be recognized at all, but also a character whose certainty factor (reliability) at the time of recognition does not reach a preset value (the certainty factor). Low characters).

In the third embodiment, the first embodiment
Although the dictionary registration for the unrecognizable character is performed on the premise of the embodiment (the one that starts the prediction operation when the recognition becomes impossible), the second embodiment (when the recognition becomes unrecognizable, It is also possible to perform dictionary registration of unrecognizable characters on the premise that a prediction operation is started when a predetermined condition is satisfied.

The method described in each of the above-described embodiments includes, as programs that can be executed by a computer, 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.

[0060]

As described above, according to the present invention, when an input character pattern becomes unrecognizable, a prediction operation is started using a character string input so far as a base character string. If the input character pattern can be recognized, the character input can be continued, and if the input character pattern cannot be recognized, a prediction candidate can be obtained to complement the character. As a result, even if the user does not explicitly instruct the start of input prediction, a candidate for a predicted character string can be obtained as needed, and a character string can be input efficiently.

When the input character pattern becomes unrecognizable and the character string input so far satisfies a predetermined condition (the number of characters set in advance), the character string based on the input character string is used. By starting the prediction operation, if the recognition becomes impossible, the problem that the prediction candidates cannot be narrowed down and the large number of prediction candidates are displayed because the number of input characters at that time is small is eliminated. Can be. Thus, the user can efficiently input a character string while supplementing characters by obtaining a certain candidate for a predicted character string narrowed down for presentation.

Further, since a character corresponding to an unrecognizable character is obtained from the predicted character string selected by the user, and the character is registered in the dictionary, the character which has become unrecognizable is recognized from the next time. Therefore, the input efficiency can be further improved.

[Brief description of the drawings]

FIG. 1 is a block diagram showing a configuration of a character string prediction device of the present invention.

FIG. 2 is a flowchart showing a flow of processing from character input to output of a predicted character string in the first embodiment of the present invention.

FIG. 3 is a flowchart showing a flow of processing from character input to output of a predicted character string in a second embodiment of the present invention.

FIG. 4 is a flowchart showing a flow of processing from character input to output of a predicted character string in a third embodiment of the present invention.

FIG. 5 is a specific example for explaining input prediction in the first embodiment.

FIG. 6 is a specific example for explaining input prediction in the second embodiment.

FIG. 7 is a specific example for describing input prediction in the third embodiment.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 101 ... Input device 102 ... Character input part 103 ... Pattern recognition part 104 ... Pattern recognition dictionary 105 ... Pattern learning dictionary 106 ... Input buffer 107 ... Prediction control part 108 ... Prediction judgment part 109 ... Dictionary search part 110 ... Recognition learning part 111 ... Word dictionary 112 candidate buffer 113 output control unit 114 output buffer 115 output device 1 input frame 2 recognition result display unit 3 prediction candidate display unit 4 recognition candidate display unit

Claims (8)

[Claims] 1. A character string predicting apparatus for predicting a character string that follows based on an input character string, comprising: input means for inputting a character pattern; and a recognition dictionary storing information necessary for pattern recognition. A recognition unit that performs a recognition process on the character pattern input by the input unit with reference to the recognition dictionary, and outputs a recognition candidate character corresponding to the input character pattern; An input buffer for sequentially storing the input characters, a word dictionary storing information of each word, and a character string stored in the input buffer when the input character pattern is processed as an unrecognizable character by the recognition means. Prediction control means for searching the word dictionary for a character string having the character string at the beginning, and predicting a character string obtained by the prediction control means. String predicting apparatus characterized by comprising a candidate display means for displaying as a candidate for string. 2. The prediction control means determines whether a character string stored in the input buffer satisfies a predetermined condition when an input character pattern is processed as an unrecognizable character by the recognition means. 2. The character string predicting apparatus according to claim 1, wherein when a predetermined condition is satisfied, a character string having the character string as a head portion is searched from the word dictionary. 3. A selecting means for selecting a candidate of a predicted character string displayed by the candidate displaying means, and a character corresponding to the unrecognizable character is specified from the predicted character string selected by the selecting means. 2. The character string prediction device according to claim 1, further comprising a recognition learning unit that associates the set character with the input character pattern processed as the unrecognizable character and registers the character in the recognition dictionary. 4. The character string prediction device according to claim 1, wherein the unrecognizable character is a character pattern that does not satisfy a certainty factor set in pattern recognition. 5. The character string prediction device according to claim 1, wherein the predetermined condition is a preset number of characters. 6. A character string prediction method for predicting a character string that follows based on an input character string, comprising the steps of: recognizing an input character pattern by referring to a recognition dictionary; Are sequentially stored in the input buffer as input characters, and when the input character pattern is processed as an unrecognizable character, based on the character string stored in the input buffer,
A character string prediction method comprising: searching a word dictionary for a character string having the character string at the beginning; and displaying the character string obtained by the search as a candidate for a predicted character string. 7. A character string prediction method for predicting a subsequent character string based on an input character string, wherein the input character pattern is recognized by referring to a recognition dictionary, and the character pattern is obtained by the recognition processing. Are sequentially stored in the input buffer as input characters, and when the input character pattern is processed as an unrecognizable character, it is determined whether or not the character string stored in the input buffer satisfies a predetermined condition. A character string that searches the word dictionary for a character string having the character string at the beginning when the predetermined condition is satisfied, and displays the character string obtained by the search as a candidate for a predicted character string. Column prediction method. 8. A character string prediction method for predicting a subsequent character string based on an input character string, wherein the input character pattern is recognized by referring to a recognition dictionary. Are sequentially stored in the input buffer as input characters, and when the input character pattern is processed as an unrecognizable character, based on the character string stored in the input buffer,
A character string having the character string at the beginning is searched from the word dictionary, and the character string obtained by this search is displayed as a candidate for a predicted character string. When the displayed predicted character string is selected, the Identify the characters that correspond to the unrecognizable characters from the predicted character string,
A character string prediction method, wherein the specified character is associated with the input character pattern processed as the unrecognizable character and registered in the recognition dictionary.