JPH08255156A - Method and device for predicting character string, and document preparation device - Google Patents

Abstract

PURPOSE: To predict a character string following a KANA(Japanese syllabary) - KANJI(Chinese character) mixed character string which are inputted for the 1st time by narrowing down character strings. CONSTITUTION: A dictionary retrieval part 3 performs retrieval from a word dictionary 4 by using the number of characters, obtained by adding the number αof added characters to the number (n) of characters in a character string inputted from an input part 1 to an input buffer 2, as the upper-limit value of the number of characters of a predicted character string candidate. Namely, the dictionary retrieval part 3 retrieves a character string, which has the character string as the head part among word consisting of HIRAGANA(cursive form of Japanese syllabary), KANJI or KANJI-mixed HIRAGANA has the total number (n+α) of characters as the upper-limit value, as the predicted character string candidate. Therefore, only predicted character string candidates which consists of (n)-(n+α) characters in total are limited and retrieved, and those predicted character strings are displayed at a display part 6 by a character string output part 5.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a character string prediction method for predicting a character string to be subsequently input by the user based on the character string input by the user, and a character string using this prediction method. The present invention relates to a prediction device and a document creation device equipped with this character string prediction device.

[0002]

2. Description of the Related Art In a conventional document creation apparatus, in order to save the effort of character input, a long character string is registered as a short character string and this is treated as a user dictionary for kana-kanji conversion, so that the user can Some have adopted a method of inputting a long character string by simply inputting a short character string. However, in this method, since the character string must be registered in advance, it is actually difficult to adopt the above method for every character string.

A method called a predictive input has been conceived for this purpose. For example, as reported in the literature (Proceedings of the 48th National Conference of the Information Processing Society of Japan 4J-11 et al.) When several characters have been entered, the word dictionary is searched using these several characters as keys, and the character string (word) that starts with these several characters is extracted and displayed. A method has been proposed in which the device predicts the sequence to save the user's input work.

However, although the above-mentioned document describes the prediction of katakana character strings (and particles), it does not consider character strings containing kanji and hiragana. Moreover, in the Japanese document creation device, considering that the hiragana input or kanji mixed hiragana input is the mainstream, the device predicts and inputs the character string following the first input character string such as hiragana or kanji. Development is required. Therefore, when predicting a hiragana or kana-kana mixed kana character string, it is indispensable to solve the following problems.

That is, when the prediction is started at the time when one hiragana character is input, there are many candidate character strings, and if all of these predicted character string candidates are retrieved from the dictionary and output, It is difficult to secure the display space, and it takes time and time to select the predicted character string desired by the user. Therefore, it is necessary to develop a method of narrowing down the number of predicted character string candidates retrieved from the dictionary without dropping the character string desired by the user. In addition, even if the number of predicted character string candidates to be searched is narrowed down, if there are multiple candidates, a measure is rearranged so that the output of the predicted character string desired by the user is prioritized,
Further, it is required to facilitate the selection of the predicted character string desired by the user.

[0006]

As described above, in the character string prediction method in which the device predicts and inputs a character string that follows from a short character string that is initially input, at present, a method for predicting a katakana character string is used. Only proposed. Moreover, in the proposed method, if there are many candidates for the character string that follows the input character string, the method to narrow down the predicted character string that is actually searched and output, or if there are multiple predicted character strings even after narrowing down, Since there is no measure for rationally rearranging the output priority of character strings and the prediction method for hiragana character strings mixed with hiragana and kanji, the character string prediction method is applied to the Japanese document creation device as it is. However, there was a problem that it was not practical.

Therefore, the present invention has been made in consideration of the above circumstances. The number of character strings following the first input kana-kanji character string is narrowed down for prediction, and there are a plurality of predicted character strings. A character string prediction method capable of rearranging so that the output priority of the character string intended by the user becomes high, a character string prediction device using this method, and a document equipped with this character string prediction device It is intended to provide a producing device.

[0008]

According to a first aspect of the present invention, a character string having an input character string at a head portion is searched from a character string stored in a dictionary, and the input character string is obtained. In a character string prediction method for predicting a subsequent character string and predicting a character string following the input character string in a character string prediction device for outputting the character string retrieved from the dictionary as a predicted character string, the input character string Number of characters added to n
(N + α) obtained by adding is used as the upper limit of the number of characters that the predicted character string has, and a method of searching the dictionary for a predicted character string having the number of characters equal to or smaller than the upper limit value is employed.

The invention of claim 2 employs a method in which the added character number α is a variable having a functional relationship with the character number n of the input character string.

The invention of claim 3 employs a method in which the additional character number α is a preset fixed value.

According to a fourth aspect of the present invention, every time the predicted character string is output and is fixed by the user, the character number m of the fixed predicted character string and the input character number n at this time are collected, and the input character number is collected. statistically processing the relationship of the number of characters m to n,
A method is adopted in which the number of added characters α is obtained from the result of this statistical processing, and the predicted character string is searched from the dictionary using the obtained number of added characters α.

According to a fifth aspect of the present invention, when there are a plurality of predicted character strings retrieved from the dictionary, the retrieved predicted character strings are rearranged in the output priority order based on one or a plurality of criteria. Has been adopted.

The invention of claim 6 employs a method of rearranging the output priorities of the predicted character strings based on the number of characters of the predicted character strings retrieved from the dictionary.

According to a seventh aspect of the invention, the usage frequency information of the character strings stored in the dictionary is stored in this dictionary,
A method of rearranging the output priority of the predicted character strings based on the frequency of use of the predicted character strings retrieved from the dictionary is adopted.

According to an eighth aspect of the invention, each time the predicted character string is output and fixed by the user, the fixed predicted character string is collected, and then the user inputs the character string. Search for a character string having this input character string at the beginning from the dictionary and the collected definite predictive character string,
As a result, when there are a plurality of retrieved predictive character strings, the output priority of the predictive character string retrieved from the collected definite predictive character strings is set to the highest priority and the output priority order of these predictive character strings is set to be higher. The output priority is rearranged.

According to a ninth aspect of the present invention, the definite predictive character string is collected by inputting an input character string forming a head portion of the definite predictive character string and a dictionary address of the dictionary accommodating the definite predictive character string. The method that is performed by collecting in the learning table is adopted.

According to a tenth aspect of the present invention, the character string following the input character string is predicted by searching the character strings stored in the dictionary for a character string having the input character string at the beginning. In a character string predicting method for predicting a character string following the input character string in a character string predicting device that outputs a character string retrieved from the dictionary as a predictive character string, a user inputs a character completion symbol together with a character. Then, a method of obtaining a predicted character string by searching the dictionary for a character string having the same character at the same position as the input character and complementing the character at the position of the input character completion symbol is adopted. are doing.

The invention of claim 11 employs a method of outputting the input character complement symbol as it is without searching the predicted character string when the input character is only the character complement symbol. ing.

According to a twelfth aspect of the present invention, there is adopted a method in which a plurality of types of the character complement symbols are used and the type of the character to be complemented at this position is specified according to the type of the input character complement symbol.

According to a thirteenth aspect of the present invention, when there are a plurality of retrieved predicted character strings, a method of rearranging the retrieved predicted character strings in the order of output priority based on one or a plurality of criteria is adopted. are doing.

According to a fourteenth aspect of the present invention, a fixed predictive character string is collected every time the predictive character string is output and confirmed by the user, and when the character string is input by the user, the input character string is input. A character string having a column at the beginning is searched from the dictionary and the collected definite predictive character string, and as a result,
When there are a plurality of retrieved predicted character strings, the output priority order of these predicted character strings is set so that the output priority order of the predicted character string retrieved from the collected definite predicted character strings is the highest priority. The method of rearranging is adopted.

According to a fifteenth aspect of the invention, the character string stored in the dictionary is hiragana, kanji, or a hiragana character string mixed with kanji.

According to a sixteenth aspect of the invention, when the input character string is a character string that does not include the character completion symbol,
A method of obtaining a predicted character string by the character string prediction method according to any one of claims 1 to 9 is adopted.

According to a seventeenth aspect of the present invention, the character string following the input character string is predicted by searching the character strings stored in the dictionary for a character string having the input character string at the beginning. In a character string prediction device that outputs a character string retrieved from the dictionary as a predicted character string, a counting unit that counts the number of characters in the character string, a numerical value supply unit that supplies the added character number α, and a numerical value supply unit that supplies the numerical value The addition character number α to be added to the character number n of the input character string counted by the counting means, and (n + α) obtained from this addition means are set as the upper limit value of the number of characters of the predicted character string. Retrieval means for retrieving a predictive character string consisting of the number of characters equal to or less than the upper limit value from the dictionary, and display means for displaying the predictive character string retrieved by the retrieving means,
The display unit has a configuration including a selection unit that allows the user to select an arbitrary character string from among the predicted character strings displayed by the display unit as a confirmed predicted character.

According to the eighteenth aspect of the present invention, the additional character number α supplied from the numerical value supplying means is a variable having a functional relationship with the character number n of the input character string counted by the counting means. .

According to the nineteenth aspect of the present invention, the additional character number α supplied from the numerical value supply means is a fixed value preset in this means.

According to a twentieth aspect of the invention, the number of characters m obtained by counting the number of characters of the confirmed predictive character string by the counting means every time the predictive character string is confirmed by the selecting means, and the counting at this time. Collecting means for collecting the input character number n already counted by the means, calculating means for statistically processing the relationship of the character number m to the input character number n collected by the collecting means, and the statistical processing result by the calculating means. An addition value updating means for obtaining the addition character number α is provided, and the addition value supplying means is configured to supply the addition character number α obtained by the addition value updating means to the searching means.

According to a twenty-first aspect of the present invention, when there are a plurality of predicted character strings retrieved from the dictionary by the retrieval means,
There is provided a rearranging means for rearranging the retrieved predicted character strings in the output priority order based on one or a plurality of criteria and then outputting the rearranged character strings to the display means.

According to a twenty-second aspect of the present invention, the rearranging means arranges the output priority order of the retrieved predicted character strings on the basis of the number of characters obtained by counting the predicted character strings retrieved by the searching means by the counting means. It has the composition to change.

The invention of claim 23 is provided with a dictionary accommodating the use frequency information of the character string, and the rearranging means is searched based on the use frequency of the predicted character string searched from the dictionary by the searching means. The output priority of the predicted character string is rearranged.

According to a twenty-fourth aspect of the present invention, the learning means collects the confirmed predictive character string every time the predictive character string is confirmed by the selecting means, and the search is performed when the character string is input by the user. The means searches the dictionary and the learning means for a character string having the input character string at the beginning, and as a result, when there are a plurality of retrieved predicted character strings, the rearranging means uses the learning means. It has a configuration in which the retrieved predicted character strings are rearranged so as to be output with the highest priority with respect to other predicted character strings.

According to a twenty-fifth aspect of the present invention, the learning means links the character string forming the head portion of the definite predictive character string and the dictionary address of the dictionary accommodating the definite predictive character string to the character string. When the same character string as the input character string is searched from the learning means, the search means searches the dictionary for a character string of the dictionary address collected by linking to the character string, and performs this search. The character string is a predicted character string retrieved from the learning means.

According to a twenty-sixth aspect of the present invention, the character string following the input character string is predicted by searching the character string stored in the dictionary for the character string having the input character string at the beginning. In a character string prediction device that outputs a character string retrieved from the dictionary as a predicted character string, input means for inputting a character complement symbol together with a character and the same character at the same position as the character input from this input means Retrieval means for retrieving from the dictionary a character string which has the character complemented at the position of the input character complement symbol, display means for displaying the predicted character string retrieved by the retrieving means, and this display And a selecting unit that allows the user to select an arbitrary character string from among the predicted character strings displayed by the means as the fixed predicted character.

According to a twenty-seventh aspect of the present invention, when only the character complement symbol is input by the input means, the search means does not search for a predicted character string and the input character complement symbol is directly displayed on the display means. It has a configuration for outputting to.

According to a twenty-eighth aspect of the present invention, there are a plurality of types of the character completion symbols input by the input means, and the search means has a specific type at the position of the character completion symbol according to the type of the character completion symbol. It has a structure that complements characters.

According to a twenty-ninth aspect of the present invention, when there are a plurality of predicted character strings retrieved from the dictionary by the retrieval means,
There is provided a rearranging means for rearranging the retrieved predicted character strings in the output priority order based on one or a plurality of criteria and then outputting the rearranged character strings to the display means.

According to a thirtieth aspect of the present invention, the learning means for collecting the determined predicted character string each time the predictive character string is fixed by the selecting means, and the search when the character string is input by the user The means searches the dictionary and the learning means for a character string having the input character string at the beginning, and as a result, when there are a plurality of retrieved predicted character strings, the rearranging means uses the learning means. It has a configuration in which the retrieved predicted character strings are rearranged so as to be output with the highest priority with respect to other predicted character strings.

According to a thirty-first aspect of the present invention, the character strings stored in the dictionary are hiragana, kanji or hiragana character strings mixed with kanji.

According to a thirty-second aspect of the invention, there is provided judging means for judging whether or not the character string inputted by the input means is a character string which does not include the character complement symbol. When the string is determined to be a character string that does not include the character completion symbol, the character string is defined as a character string.
It has a configuration for obtaining a predicted character string by any one of the character string prediction devices.

A thirty-third aspect of the present invention is a document creating apparatus for creating a document by inputting a character string, the thirteenth to third aspects.
2 is equipped with the character string prediction device described above, and the predicted character string determined by the character string prediction device is input into the document.

[0041]

According to the character string prediction method of the invention of claim 1,
(N + α), which is obtained by adding the number of additional characters α to the number of characters n of the input character string, is set as the upper limit of the number of characters of the predicted character string, and the predicted character string having the number of characters equal to or less than the upper limit value is searched from the dictionary.

In the character string prediction method according to the second aspect of the present invention, the added character number α is a variable having a functional relationship with the character number n of the input character string.

In the character string prediction method according to the third aspect of the present invention, the additional character number α is a preset fixed value.

According to the character string prediction method of the invention of claim 4, each time the predicted character string is output and is fixed by the user, the fixed character number m of the predicted character string and the input character number at this time are set. n is collected, the relationship between the number of input characters n and the number of characters m is statistically processed, the additional character number α is obtained from the result of the statistical processing, and the predicted character string is searched from the dictionary using the obtained additional character number α. .

According to the character string prediction method of the invention of claim 5, when there are a plurality of predicted character strings retrieved from the dictionary, the retrieved predicted character strings are based on one or a plurality of criteria. Are sorted in order of output priority.

In the character string prediction method according to the sixth aspect of the present invention, the output priority order of these predicted character strings is rearranged on the basis of the number of characters of the predicted character strings retrieved from the dictionary.

According to the character string prediction method of the invention of claim 7, the usage frequency information of the character string stored in the dictionary is stored in this dictionary, and the predicted character string retrieved from the dictionary is stored. The output priorities of the predicted character strings are rearranged based on the frequency of use.

According to the character string prediction method of the invention of claim 8, each time the predicted character string is output and is fixed by the user, the fixed predicted character string is collected, and then the user predicts it. When a character string is input, a character string having this input character string at the beginning is searched from the dictionary and the collected definite predictive character string, and as a result, there are a plurality of predicted character strings searched. In this case, the output priority order of these predicted character strings is rearranged so that the output priority order of the predicted character strings retrieved from the collected definite predictive character strings is the highest priority.

In the character string predicting method of the present invention, the definite predictive character string is collected by accommodating the input character string forming the head portion of the definite predictive character string and the definite predictive character string. This is done by collecting in the learning table the dictionary addresses of said dictionaries.

According to the character string prediction method of the tenth aspect of the invention, when the user inputs the character complement symbol together with the character,
A predicted character string is obtained by searching the dictionary for a character string having the same character at the same position as the input character and complementing the character at the position of the input character completion symbol.

In the character string prediction method according to the invention of claim 11, when the input character is only the character completion symbol, the predicted character string is not searched and the input character completion symbol is not searched. Is output as is.

In the character string prediction method according to the invention of claim 12, a plurality of types of the character complement symbols are used, and the type of character to be complemented at this position is specified according to the type of the input character complement symbol. There is.

In the character string prediction method of the thirteenth aspect, when there are a plurality of retrieved predicted character strings, the retrieved predicted character string is output based on one or a plurality of criteria. Sort by priority.

In the character string prediction method of the fourteenth aspect of the present invention, a fixed predicted character string is collected every time the predicted character string is output and fixed by the user, and the character string is set by the user. When input, a character string having this input character string at the beginning is searched from the dictionary and the collected definite predictive character string, and as a result, when there are a plurality of predicted character strings searched, the The output priority order of these predicted character strings is rearranged so that the output priority order of the predicted character strings retrieved from the collected definite predictive character strings is the highest priority.

In the character string prediction method of the fifteenth aspect of the present invention, the character strings stored in the dictionary are hiragana, kanji, or hiragana character strings mixed with kanji.

In the character string prediction method of the invention of claim 16, in the case where the input character string is a character string which does not include the character completion symbol, the character string of any one of claims 1 to 9 A predicted character string is obtained by the prediction method.

In the character string prediction device of the seventeenth aspect of the present invention, the counting means counts the number of characters in the character string. The numerical value supply means supplies the additional character number α. The adding means adds the added character number α supplied from the numerical value supplying means to the character number n of the input character string counted by the counting means. The searching means sets (n + α) obtained from the adding means as an upper limit of the number of characters that the predicted character string has, and searches the dictionary for a predicted character string having a number of characters equal to or less than the upper limit value. The display means displays the predicted character string retrieved by the retrieval means. The selection unit causes the user to select an arbitrary character string as the confirmed predicted character from the predicted character strings displayed by the display unit.

In the character string predicting apparatus of the eighteenth aspect of the present invention, the additional character number α supplied from the numerical value supplying means is a variable having a functional relationship with the character number n of the input character string counted by the counting means. .

In the character string predicting device of the nineteenth aspect of the present invention, the additional character number α supplied from the numerical value supply means is a fixed value preset in this means.

In the character string predicting apparatus of the invention of claim 20, the collecting means obtains by counting the number of characters of the confirmed predictive character string by the counting means every time the predicting character string is confirmed by the selecting means. The number of characters m and the number of input characters n already counted by the counting means at this time are collected. The calculating means statistically processes the relationship between the number of input characters n collected by the collecting means and the number of characters m. The addition value updating means obtains the addition character number α from the statistical processing result by the calculating means. The additional value supplying means supplies the additional character number α obtained by the additional value updating means to the searching means.

In the character string predicting apparatus of the twenty-first aspect of the invention, when there is a plurality of predictive character strings retrieved from the dictionary by the retrieving means, the rearranging means selects one of the retrieved predictive character strings. The data is sorted in the output priority order based on a plurality of criteria and then output to the display means.

In the character string predicting apparatus of the twenty-second aspect of the present invention, the rearrangement means searches the predicted character string based on the number of characters counted by the counting means for the predicted character string searched by the searching means. Rearrange the output priority of.

In the character string predicting apparatus of the twenty-third aspect of the invention, the dictionary stores the frequency-of-use information of the character string. The rearranging unit rearranges the output priority order of the retrieved predicted character strings based on the frequency of use of the predicted character strings retrieved from the dictionary by the searching unit.

In the character string predicting device of the twenty-fourth aspect of the invention, the learning means collects the confirmed predicted character string every time the selecting means determines the predicted character string. When the user inputs a character string, the search means searches the dictionary and the learning means for a character string having the input character string at the head portion, and as a result, a plurality of searched predicted character strings are searched. If there is, the rearrangement unit performs rearrangement so that the predicted character string retrieved by the learning unit is output with the highest priority with respect to other predicted character strings.

In the character string predicting apparatus of the twenty-fifth aspect of the present invention, the learning means stores the character string forming the head portion of the definite predictive character string and the dictionary address of the dictionary containing the definite predictive character string. When the search means searches the learning means for the same character string as the input character string, the character string of the dictionary address collected by linking to the character string is collected from the dictionary. A search is performed, and this retrieved character string is used as the predicted character string retrieved from the learning means.

In the character string predicting device of the twenty-sixth aspect of the invention, the input means inputs the character complement symbol together with the character.
The search means searches the dictionary for a character string having the same character at the same position as the character input from the input means and complementing the character at the position of the input character completion symbol. The display means displays the predicted character string retrieved by the retrieval means. The selection unit causes the user to select an arbitrary character string as the confirmed predicted character from the predicted character strings displayed by the display unit.

In the character string predicting device of the twenty-seventh aspect of the present invention, when only the character complement symbol is input by the input means, the searching means does not search for the predicted character string, but the input character complement. The symbol is output as it is to the display means.

In the character string prediction device of the twenty-eighth aspect of the present invention, there are a plurality of types of the character completion symbols input by the input means, and the retrieval means generates the character completion symbols according to the type of the character completion symbols. Completing a particular type of character at a position.

In the character string prediction apparatus of the 29th aspect of the invention, when there is a plurality of predicted character strings retrieved from the dictionary by the retrieving means, the rearranging means may use one of the retrieved predictive character strings. The data is sorted in the output priority order based on a plurality of criteria and then output to the display means.

In the character string predicting apparatus of the thirtieth aspect, the learning means collects the confirmed predicted character string every time the selecting means determines the predicted character string. When the user inputs a character string, the search means searches the dictionary and the learning means for a character string having the input character string at the head portion, and as a result, a plurality of searched predicted character strings are searched. If there is, the rearrangement unit performs rearrangement so that the predicted character string retrieved by the learning unit is output with the highest priority with respect to other predicted character strings.

In the character string predicting apparatus of the thirty-first aspect of the invention, the character strings stored in the dictionary are hiragana, kanji, or hiragana character strings mixed with kanji.

In the character string predicting device of the thirty-second aspect of the invention, the judging means judges whether or not the character string input by the input means is a character string that does not include the character completion symbol. When the input character string is determined to be a character string that does not include the character complement symbol by the determination means, a predicted character string is obtained by the character string prediction device according to any one of claims 17 to 25.

A document creating apparatus according to a thirty-third aspect of the invention is equipped with the character string predicting apparatus according to any one of the seventeenth to thirty-second aspects, and a predicted character string determined by the character string predicting apparatus is input into the document. To do.

[0074]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of the present invention will be described below with reference to the drawings. FIG. 1 is a block diagram showing an embodiment of a character string prediction device of the present invention using the character string prediction method of the present invention. The character string prediction device includes an input unit 1, an input buffer 2, a dictionary search unit 3, a word dictionary 4, a character string output unit 5, and a display unit 6. Here, the function of each block described above will be described. The input unit 1 inputs a character string by recognizing a handwritten character from coordinate data obtained from a handwriting tablet or the like, or includes a keyboard or the like. The input buffer 2 stores the character string input from the input unit 1 until it becomes a prediction target. The dictionary search unit 3 searches the word dictionary 4 for a predicted character string having the input character string at the beginning. The word dictionary 4 contains a character string composed of hiragana, kanji, and hiragana mixed with kanji. The character string output unit 5 displays the predicted character string and the like on the screen of the display unit 6. The display unit 6 is LC
It consists of D, CRT, etc., and displays input character strings, predicted character strings, and the like.

Next, the operation of this embodiment will be described. When the user is inputting a character (string) from the input unit 1 and wants the device to predict a subsequent character string, the user presses a prediction key preset in the input unit 1. The input unit 1 temporarily stores the input character string in the input buffer 2, but when the prediction key is pressed, the character string input in the input buffer 2 until then is input character string The same) is output to the dictionary search unit 3. For example, if the user is "to"
When the prediction key is pressed at the time of inputting, "to" is input from the input buffer 2 as an input character string, and the dictionary search unit 3
If the prediction key is pressed when the user inputs "Tori", "Tori" is output from the input buffer 2 to the dictionary retrieval unit 3 as an input character string.

The dictionary retrieval unit 3 has the input character string from the input buffer 2 at the head portion, and the sum of the character number n of this input character string and the preset addition character number α (n + α).
In the following, a character string consisting of n or more characters is searched from the word dictionary 4 as a predicted character string candidate, and the obtained predicted character string candidate is sent to the character string output unit 5. As a result, the dictionary search unit 3 narrows down the number of predicted character string candidates to be searched by determining the upper limit of the number of characters constituting the predicted character string candidates searched from the word dictionary 4. However, the added character number α is preset in the dictionary search unit 3. Here, a concept for narrowing down the number of predicted character string candidates to be searched as described above will be described. If the user wants to predict a long character string, the first character string to input is long, and if the user wants to predict a short character string, the first character string to input is also short. Therefore, if you subtract the number of characters in the first input string from the number of characters in the predicted character string that the user intended to input, the upper limit value is found, and the character string with the number of characters exceeding this upper limit value becomes the predicted character string candidate. It seems that there is nothing. Therefore, even if the number of characters α obtained by adding a certain number to the number of characters of the character string to be input first is limited to the maximum number of characters that the predicted character string has, the predicted character string that the user intends to input is included in this. Since the probability is high, it is considered that the predicted character string intended for input will not be leaked from the candidates even with the above limitation.

Next, the character string output unit 5 displays the sent predicted character string on the display unit 6. The method of displaying the predicted character string of the character string output unit 5 is as follows: the predicted character string candidates that have been passed are displayed in a list on the display unit 6 or the order in which they are passed is given each time there is a next candidate output instruction from the user. Display sequentially. If there is a desired predicted character string displayed on the display unit 6, the user operates the input unit 1 to select the desired character string, thereby determining the predicted character string.

FIG. 2 is a flow chart showing the above-mentioned character string prediction processing procedure. The input unit 1 initializes the input buffer 2 in step 201, and then stores the character input in step 202 in the input buffer 2 in step 203. The input unit 1 determines whether or not the prediction key is pressed in step 204 and the character string in the input buffer 2 is the prediction target. If not, the process returns to step 202 to repeat the input of characters, If not, go to step 205. The dictionary search unit 3 searches the word dictionary 4 for a predicted character string candidate having the character string sent from the input buffer 2 at the head in step 205 within the above-described character number limit, and searches the predicted character string candidate for the character string. It is sent to the column output unit 5. The character string output unit 5 displays a list of predicted character string candidates passed to the display unit 6 in step 206. At step 207, the user selects a desired predicted character string candidate displayed on the display unit 6. However, if there is no desired character string, a new character is input from the input unit 1.
The input unit 1 determines in step 208 whether or not the predicted character string is confirmed, and if not confirmed, in step 202
Returning to step 2, the input of characters and prediction are repeated, and if confirmed, the process proceeds to step 209. The character string output unit 5 displays only the confirmed predicted character string on the display unit 6. It should be noted that the input unit 1 determines that the predicted character string is not finalized when the user does not select the predicted character string and inputs a new character.

FIG. 3 is a diagram showing an example of predicted character string candidates retrieved from the word dictionary 4 by the dictionary retrieval unit 3 shown in FIG. In the prediction, if the number of added characters α = 2, for example, (the number of input characters n + 2) becomes the upper limit of the number of characters of the predicted character string candidate. As shown in FIG. 3A, when the user inputs “Asa”, for example, “Asa-o (morning glory)”,
A character string of 2 to 4 characters such as “Asahi (Asahi)” and “Asa (Asahi)” is searched as a predicted character string. Also, FIG.
When "Asahi" is entered as in (B), the upper limit of the number of characters in the predicted character string changes to 5 characters, so a 5-character string such as "Asahikawa (Asahikawa)" is predicted. It will be newly added to the character string candidates.

According to the present embodiment, the dictionary search unit 3 searches the word dictionary 4 for a predictive character string candidate having the upper limit of the number of characters obtained by adding the number of added characters α to the number of characters n of the input character string. The number of predicted character string candidates to be input can be narrowed down without leaking the predicted character string that the user intends to input, and it is possible for the user to easily select the predicted character string that is intended and to search for it. The processing time can be improved by shortening the time. Further, when the list of predicted character string candidates is displayed on the display unit 6, the number of predicted character string candidates is narrowed down as described above, so that a space for displaying a list on the display unit 6 can be secured without difficulty. still,
In the above-described embodiment, the predicted character string candidate having the number of characters obtained by adding a certain number of characters α to the number of characters n of the first character string input by the user is searched, but α to be added to the number n of input characters is searched. As a function of the input character number n, the number of input character strings can be changed according to the number of characters in the input character string. By doing so, it is possible to more effectively narrow down the number of searches for predicted character string candidates. That is, in this case, α = f (n), and the upper limit of the number of characters of the predicted character string candidate is {n + f (n)}. Further, not only is the upper limit value of the predicted character string candidates searched by the dictionary search unit 3 set, but the lower limit value is set to a value other than n, or the correspondence table of the number of input characters and the number of characters of the predicted character string is used. The character number range of the predicted character string may be set. Further, as a method for determining that the character string in the input buffer 2 has been a prediction target, a method other than pressing the prediction key of this example may be used.

FIG. 4 is a block diagram showing another embodiment of the present invention. The character string prediction device includes an input unit 1, an input buffer 2, a dictionary search unit 3, a word dictionary 4, a character string output unit 5, a display unit 6, a prediction candidate control unit 7, and a candidate buffer 8. Here, the functions of the blocks that are not used in the device shown in FIG. 1 or have different functions among the above blocks will be described. The word dictionary 4 contains a character string consisting of hiragana, kanji, and hiragana mixed with kanji, and usage frequency information of each character string. The character string candidate control unit 7 rearranges the predicted character strings in the output priority order and then outputs them to the character string output unit 5. The display unit 6 displays an input character string, a predicted character string, and the like.

Next, the operation of this embodiment will be described with reference to the flowchart of FIG. Input unit 1 is step 50
After initializing the input buffer 2 in step 1, step 502
The character is input at, and the input character is stored in the input buffer 2 at step 503. The input unit 1 is step 504.
At, the prediction key is pressed to determine whether or not the character string in the input buffer 2 has become the prediction target. If not, the process returns to step 502 to repeat the character input. If it is a prediction target, the process proceeds to step 505. The dictionary search unit 3 searches the word dictionary 4 for a predicted character string candidate having the character string sent from the input buffer 2 at the head in step 505,
The retrieved predicted character string is stored in the candidate buffer 8. At this time, the dictionary retrieval unit 3 also extracts the usage frequency information of the retrieved predicted character string candidate from the word dictionary 4 and stores it in the candidate buffer 8. Here, the dictionary search unit 3 has a character string having the input character string from the input buffer 2 at the head portion and having a character number smaller than the sum (n + α) of the character number n of this input character string and the preset additional character number α. Is searched as a predicted character string candidate from the word dictionary 4. However, α = f (n) may be set.

The predictive candidate control unit 7 rearranges the predictive character string candidates stored in the candidate buffer 8 in step 506 in the output priority order according to the usage frequency information, and then outputs the rearranged predictive character string candidates as a character string. Send to Part 5. The character string output unit 5 displays the predicted character string candidates sent in step 507 on the display unit 6. The method of displaying the predicted character string candidates of the character string output unit 5 is such that the predicted character string candidate having the first output priority is displayed first, and if the next candidate selection instruction is given from the input unit 1, the second character prediction is performed. Display character string candidates. The same applies to the display of the predicted character string candidates in the third and subsequent ranks. Alternatively, the character string output unit 5 may display a list of the passed predicted character string candidates on the display unit 6 in order of output priority. User step 5
At 08, the input unit 1 is operated to select a desired predictive character string candidate. As a result, the input unit 1 determines whether or not the predicted character string is fixed in step 509. If the predicted character string is not fixed, the process returns to step 502 to repeat the character input and prediction, and if the character string is fixed, the process proceeds to step 510. move on. The character string output unit 5 displays the predicted character string confirmed in step 510 on the display unit 6.

According to the present embodiment, the predictive character string candidates retrieved by limiting the number of characters by the dictionary retrieval unit 3 can be rearranged in the output priority order and displayed on the display unit 6,
It is possible to facilitate the selection of the predicted character string desired by the user, increase the probability of displaying the predicted character string desired by the user in the first order, and improve the operability of the device. The prediction candidate control unit 7 may rearrange the predicted character string candidates having a larger number of characters so that the priority order thereof is higher. The reason for this is that if you want to predict the characters that the user will input, it is difficult to enter a word in a long character string, so just enter the first few characters and let the device predict the rest. Is the general idea. According to this idea, it is considered that the larger the number of characters among the predicted character string candidates, the higher the possibility that the character string is intended for input by the user. Further, the method of determining the priority order based on the number of characters of the predicted character string candidate and the method of determining based on the usage frequency information stored in the word dictionary 4 may be used together. In this case, priority is given to the determination of the output priority order based on the usage frequency information, and in the case of predicted character string candidates having the same usage frequency, the priority order is further determined based on the number of characters.

FIG. 6 is a block diagram showing another embodiment of the present invention. The character string prediction device includes an input unit 1, an input buffer 2, a dictionary search unit 3, a word dictionary 4, a character string output unit 5, a display unit 6, a prediction candidate control unit 7, a candidate buffer 8, an input learning unit 9, and a learning table 10. Made of. Here, among the above-mentioned blocks, the functions of blocks not used in the apparatus shown in FIG. 4 will be described. When the predicted character string is confirmed by the user, the input learning unit 9 stores in the learning table 10 the character string input from the input unit 1 serving as a prediction key together with the dictionary address of the confirmed character string. The learning table 10 stores the input character string and the dictionary address of the confirmed character string.

Next, the operation of this embodiment will be described with reference to the flowchart of FIG. The input unit 1 is step 70
After initializing the input buffer 2 in step 1, step 702
The character is input at, and the character input at step 703 is stored in the input buffer 2. The input unit 1 is step 704.
The prediction key is pressed at, and it is determined whether the character string in the input buffer 2 is the prediction target. If the prediction string is not the prediction target, the process returns to step 702, and the character input is repeated to become the prediction target. In that case, the process proceeds to step 705. The dictionary search unit 3 searches the word dictionary 4 for the predicted character string candidates having the character string sent from the input buffer 2 at the head in step 705.
Then, the predicted character string candidates retrieved from the learning table 10 and retrieved from the word dictionary 4 are stored in the candidate buffer 8. At this time, the dictionary retrieval unit 3 also extracts the usage frequency information of the retrieved predicted character string candidate from the word dictionary 4 and stores it in the candidate buffer 8. Here, the dictionary search unit 3 predicts from the word dictionary 4 a character string having an input character string at the beginning and having a character number smaller than the sum (n + α) of the character number n of this input character string and a preset character number α. Search as a character string candidate. However, α = f (n) may be set.

Next, the dictionary retrieval unit 3 determines in step 706 whether or not the character string is retrieved from the learning table 10. If it is not retrieved, the process jumps to step 708, and if it is retrieved, the process proceeds to step 707. The dictionary search unit 3 searches the word dictionary 4 for a fixed character string (here, the same as the predicted character string) that includes the character string searched for from the learning table 10 at step 707, and obtains the predicted character string candidate. Is added to the candidate buffer 8. The prediction candidate control unit 7 rearranges the prediction character string candidates stored in the candidate buffer 8 in step 708 according to the usage frequency information and the number of characters of the prediction character string, and then rearranges the rearranged prediction character string candidates in the character string output unit. Send to 5. However, when there is a predicted character string candidate retrieved from the word dictionary 4 based on the character string retrieved from the learning table 10 when rearranging by the predictive candidate control unit 7, this predicted character string candidate Shall be the highest priority and shall be sorted in the order of output priority. The reason for this is that it is highly likely that the predicted character string that the user has predicted by the device and confirmed is input by the user many times.

The character string output unit 5 displays the predicted character string candidates sent in step 709 on the display unit 6 to prompt the user to select the predicted character string. At this time, the method of displaying the predicted character string candidates of the character string output unit 5 is such that when the predicted character string candidate having the first output priority is first displayed and the next candidate selection instruction is given from the input unit 1, Display the predicted character string candidates for the rank. The same applies to the display of the predicted character string candidates in the third and subsequent ranks. Alternatively, the character string output unit 5 may display a list of the passed predicted character string candidates on the display unit 6 in order of output priority. The user operates the input unit 1 in step 710 to select a desired predictive character string candidate. As a result, the input unit 1 determines in step 711 whether or not the predicted character string is fixed. If not, the process returns to step 702 to repeat the input and prediction of the character, and if confirmed, to step 712. move on. The input learning unit 9 is step 712.
It is determined whether or not the prediction candidate confirmed in (3) is a character string directly searched from the word dictionary 4 without going through the learning table 10. If not, the process jumps to step 714, and if so, the process proceeds to step 713. The input learning unit 9 registers the dictionary address in the word dictionary 4 of the predicted character string selected and confirmed as the input character string in the learning table 10 as shown in FIG. Thereafter, the character string output unit 5 displays the confirmed predicted character string in step 714 on the display unit 6.
To be displayed.

In the example of FIG. 8, when the character string stored in the input buffer 2 is "Asame", the input character string "Asame" and the predicted character string "Asameshimae (Amae Mae) ) ”Word dictionary 4 dictionary address“ 0x
000a ”is registered in the learning table 10.

FIG. 8 is a diagram showing a storage example of the learning table 10 as described above. The learning table 10 stores an input character string (learning character string) and a dictionary address of the word dictionary 4 in which a predicted character string including the input character string at the head and once determined by the user is stored. . For example, when the input character string is "Asame" and the definite predictive character string at this time is "Asameshimae (Amae Mae)", the dictionary address is "0x000a". Naturally, "0x000a" in the word dictionary 4 shown in FIG.
"Asameshimae (Amae Mae)" is stored at the dictionary address of.

Next, a specific example of the character string prediction process described above will be described. First, when "Asa" is input from the input unit 1 without learning, a character string having "Asa" at the head portion is output as a predicted character string candidate as shown in FIG. Further, when the third character "me" is additionally input as shown in FIG. 10B, the predictive character string candidates are "asameshimae (before breakfast)" and "asameshi (breakfast)".
Is narrowed down to. Here, "Asameshimae (before breakfast)"
When the predicted character string candidate is confirmed, the learning table 10 shown in FIG. 8 stores the input character string at that time and the dictionary address of the word dictionary 104 in which the confirmed predicted character string is stored. In the state where learning is performed in this way, FIG.
As shown in (C), when "Asa" is input from the input unit 1, "Asame" is searched from the learning table 10, and the predicted character string previously determined based on the dictionary address stored in the learning table 10 is searched. "Asameshimae (before breakfast)"
Is searched from the word dictionary 10 and added to the prediction character string candidates in the candidate buffer 8, and then the prediction candidate control unit 7 arranges the output priority of "Aseshimae (Amae Mae)" to be first. Since it is replaced and output is given priority,
Predicted character string candidates in the output priority order as shown in FIG. 10C are obtained.

According to the present embodiment, the learning table 10 collects the input character string portion of the predicted character string once determined by the user and the dictionary address in the word dictionary 4 of the predicted character string, The predicted character string candidates obtained by the subsequent prediction process are stored in the word dictionary 4 via the information in the learning table 10.
If it is searched for from the prediction candidate control unit 7, the prediction candidate control unit 7 rearranges the prediction character string candidates with the output priority order of the prediction character string candidates as the highest priority. There is a high possibility that the sequence will be output in the first order, and it is possible to easily determine the predicted character string and improve the operability of the device. As long as the capacity of the learning table 10 can be secured, the input character string and the confirmed predictive character string at that time are stored in the learning table 10, and the dictionary search unit 3 uses only the learning table 10 to predict the predicted character once. The same effect can be obtained by using a structure for searching a column.

FIG. 11 is a block diagram showing still another embodiment of the present invention. The character string prediction device comprises an input unit 1, an input buffer 2, a dictionary search unit 3, a word dictionary 4, a character string output unit 5, a display unit 6, a prediction candidate control unit 7, a candidate buffer 8 and a flag buffer 11. Here, among the above-mentioned blocks, the functions of blocks not used in the apparatus shown in FIG. 4 will be described. Flag buffer 11
Is a buffer in which a flag is set by the input unit 1 when the character complement instruction information is input from the input unit 1. Next, the operation of this embodiment will be described with reference to the flowchart of FIG. The input unit 1 initializes the input buffer 2, the candidate buffer 8, and the flag buffer 11 in step 121, and inputs characters in step 122. This time,
Input unit 1 is the character (symbol) input in step 123
Is to indicate completion of characters,
If so, jump to step 125, else go to step 124. The input unit 1 sets a flag in the flag buffer 11 in step 124, and then executes step 1
Go to 25. The input unit 1 stores the input character (symbol) in the input buffer 2 and then determines whether the character string prediction instruction is input in step 126 and the character string stored in the input buffer 2 becomes a prediction target. If not, the process returns to step 122 to repeat the input of characters, and if it becomes a prediction target, the character string in the input buffer 2 is sent to the dictionary search unit 3, and then the process proceeds to step 127. The dictionary search unit 3 determines in step 127 whether or not the flag is set in the flag buffer 11. If not, the process proceeds to step 129, and if it is set, step 128.
Proceed to. The dictionary search unit 3 searches the word dictionary 4 for a character string that has the same character in the same position as the character input in step 128 and that can be supplemented with a character in the position where the completion of the character is instructed. After storing the character string in the candidate buffer 8, the process proceeds to step 130. On the other hand, in the case of proceeding to step 129, the dictionary retrieval unit 3 stores the symbol instructing the complement of the inputted character in the candidate buffer 8 as it is, and then proceeds to step 130. The predictive candidate control unit 5 rearranges the predictive character string candidates stored in the candidate buffer 8 in step 130 in the output priority order according to the number of characters and the usage frequency stored in the word dictionary 4, and then rearranges the predictive character strings. The column candidates are sent to the character string output unit 5. The character string output unit 5 displays the predicted character string sent in step 131 on the display unit 6. The method of displaying the predicted character string candidates of the character string output unit 5 is such that the predicted character string candidate having the first output priority is displayed first, and if the next candidate selection instruction is given from the input unit 1, the second character prediction is performed. Display character string candidates. The same applies to the display of the predicted character string candidates in the third and subsequent ranks. Alternatively, the character string output unit 7 may list the passed predicted character string candidates on the display unit 8 in order of output priority. The user operates the input unit 1 in step 131 to select a desired one from the predicted character strings. As a result, the input unit 1 determines whether or not the predicted character string is fixed in step 133. If the predicted character string is not fixed, the process returns to step 122, the character input and prediction are repeated, and if the character string is fixed, the process proceeds to step 134. move on. The character string output unit 5 displays the predicted character string confirmed in step 134 on the display unit 6.

FIG. 13 is a diagram for explaining the above-described character string predicting operation as a concrete example. In the present prediction example, “◯” is a symbol for instructing character complement, and “*” is a symbol for instructing character complement by Kanji. As shown in FIG. 13 (A), when "Asa OOOO" is input from the input unit 1, the dictionary search unit 3 has "Asa" at the beginning and can supplement the characters in "O". The character string "Asameshimae" is searched from the word dictionary 4 and used as a predicted character string candidate. Further, when "Foo ◯" or "Foo *" is input from the input unit 1 as shown in FIG. 13B, "Foot" and the like are obtained as predicted character string candidates in either case. When "Foo * O" is entered as in 13 (C), "Footstep" is obtained as a predicted character string candidate. Note that FIG. 13D is a specific example in the case of proceeding to step 139 in the flowchart of FIG. 12 described above, and in the case where only a symbol for instructing character complementation such as “◯” is input from the input unit 1. Are stored in the candidate buffer 8 as they are and displayed on the display unit 6 by the character string output unit 5.

According to the present embodiment, since the complement position of the character string is clearly indicated by the user, the number of predicted character strings can be considerably narrowed down, and moreover, the character string which the user intends to input is Since the probability of prediction is extremely high, it is possible to predict a character string with extremely high accuracy, and it is possible to easily select a predicted character string. or,
Since it is possible to limit the type of character to be complemented at the place by the complement instruction symbol of the character string, it is possible to further increase the probability that the character string intended to be input by the user is predicted. In this example, the circle and the US mark are used as the complementary instruction symbols of the character string, but if a newly created dedicated complementary instruction symbol other than symbols used in other meanings is used, the flag is set in the input buffer 11. It is not necessary to distinguish the case where the complementary instruction symbol is used in another meaning by omitting the standing structure, and the structure of the device can be simplified. Further, when the complementary instruction symbol is input as in this embodiment, the dictionary search unit 3 searches for the predicted character string as described above. However, when a normal character string that does not include a complementation instruction symbol is input, a configuration is used in which the number of characters of the predicted character string is limited and the word dictionary 4 is searched as in the embodiment shown in FIG. You can also do it. In this case, one or both of the learning function shown in FIG. 6 and the additional character number learning function shown in FIG. 14 may be added, but in any case, the input character string is complemented. The input unit 1 determines whether or not the instruction symbol is included, and the character prediction method is switched according to the determination result. Further, the prediction candidate control unit 7 and the candidate buffer 8 are omitted from the configuration of this example shown in FIG.
It is also possible to employ the configuration shown in FIG. 1 in which the predicted character string retrieved by is directly output to the character string output unit 5.

FIG. 14 is a block diagram showing still another embodiment of the present invention. The character string prediction device of this example has an input unit 1,
The input buffer 2, the dictionary search unit 3, the word dictionary 4, the character string output unit 5, the display unit 6, and the additional value learning unit 12 are included. Here, the function of the added value learning unit 12 which is not used in the device shown in FIG. 1 in each of the blocks described above will be described. The additional value learning unit 12 uses α for determining the upper limit of the number of characters of the predicted character string from the number of characters of the confirmed predicted character string.
Is statistically obtained and provided to the dictionary search unit 3.

Next, the operation of this embodiment will be described with reference to the flowchart of FIG. Input section 1 is step 1
After initializing the input buffer 2 at 51, step 15
A character is input in step 2, and the input character is stored in the input buffer 2 in step 153. Input unit 1 is step 15
In step 4, the prediction key is pressed to determine whether or not the character string in the input buffer 2 is the prediction target. If not, the process returns to step 152 to repeat the input of characters, and if it is the prediction target, the step is performed. Proceed to 155. Dictionary search unit 3
Has the input character string from the input buffer 2 at the beginning in step 155, and sums the number of characters n of this input character string and the additional character number α provided from the additional value learning unit 12 (n +
A character string having a smaller number of characters than α) is searched from the word dictionary 4 as a predicted character string candidate, and the obtained predicted character string candidate is sent to the character string output unit 5. The character string output unit 5 is step 156.
A list of predicted character string candidates passed to the display unit 6 is displayed. In step 157, the user selects and confirms the desired one from the predicted character strings displayed on the display unit 6. The input unit 1 determines whether or not the predicted character string is confirmed in step 158, and if not, proceeds to step 1
By returning to 55 and performing the above-mentioned steps 152 to 158, the character input and the prediction process are repeated.
When the user selects and confirms the predicted character string, the input unit 1 receives the operation at this time and informs the character output unit 5 of the confirmed character string. When the predicted character string is fixed, the additional value learning unit 12 receives the fixed predicted character string through the character string output unit 5 in step 159, and it is the head portion of the predicted character string from the character number m of the fixed predicted character string. The number of characters n of the input character string (character string in the input buffer 2) is subtracted to obtain mn = a. A for this input character number n
Then, the average of a with respect to the number of input characters n, which has been obtained in the same manner as above, is obtained, and the average of a with respect to the number of input characters n is calculated.

Next, the most appropriate added character number α is obtained based on the calculated average value of a, and the added character number α with respect to the character number n is provided to the dictionary retrieval unit 3. As a result, the added character number α becomes a value that reflects the habit of the user. However, since nothing is initially learned, the additional character number learning unit 12 determines the typical additional character number α for the character number n by the dictionary search unit 3
Shall be provided to The character string output unit 5 is step 16
The predicted character string determined by 0 is displayed on the display unit 6.

According to the present embodiment, the predicted character for which the user has actually determined the value of the additional character number α used when the upper limit of the number of characters of the predicted character string candidates searched by the dictionary search unit 3 is determined from the input character number n. By statistically deriving from the string, it is possible to increase the probability that the predicted character string searched by the dictionary search unit 3 contains the one that the user intended to input, according to the habit of the user. At the same time, since the upper limit of the number of characters is not unnecessarily increased, the character prediction speed can be increased.

FIG. 16 is a block diagram showing still another embodiment of the present invention. This example has a configuration in which an additional character number learning function is added to the device shown in FIG. This allows
In addition to the effect of the device shown in FIG. 4, the character prediction accuracy of the dictionary search unit 3 can be increased according to the habit of the user,
In addition, the character prediction speed can be increased.

FIG. 17 is a block diagram showing still another embodiment of the present invention. This example has a configuration in which an additional character number learning function is added to the device shown in FIG. This allows
In addition to the effect of the apparatus shown in FIG. 6, the character prediction accuracy of the dictionary search unit 3 can be increased according to the habit of the user,
In addition, the character prediction speed can be increased. The present invention can also be applied to the prediction of katakana character strings in the same manner to obtain the same effect. Further, in the above-mentioned embodiment, the place written as hiragana character string is read as katakana character string, and the kana-kanji conversion is read as katakana kanji conversion.
The present invention can be easily applied to the prediction of character strings of Chinese characters and Katakana, and similar effects can be achieved. Further, by mounting the above-described character string prediction device of the present invention on the document creation device, the user can input all the character strings by only inputting a part of the beginning of the input character string. Therefore, the document creation efficiency can be improved.

[0102]

As described above, claims 1, 2, 3, 1
According to the inventions of 7, 18, and 19, the number of predicted character string candidates retrieved from the dictionary can be narrowed down to easily secure the display space of the predicted character string candidates, and the predicted character string desired by the user. Can be selected easily. According to the invention of claim 4 or 20, it is possible to adjust the upper limit number of characters of the predicted character string to be searched according to the habit of the user,
The prediction efficiency of the character string can be improved.

According to the invention of claims 5, 6, 7, 21, 22, and 23, the predicted character string candidates predicted by narrowing down the number can be rearranged in the output priority order, and the predicted character desired by the user. The operability of the device can be improved by increasing the output rank of the column.

According to the eighth, ninth, twenty-fourth and twenty-fifth aspects of the present invention, when the predicted character string once selected by the user becomes a candidate again, the predicted character string candidate is used with higher output priority. It is possible to improve the operability of the device by facilitating the selection of the predicted character string desired by the person.

According to the tenth or twenty-sixth aspect of the present invention, it is possible to further narrow down the predicted character string to be searched and increase the probability that the predicted character string desired by the user is included therein.

According to the eleventh or twenty-seventh aspect of the present invention, an existing symbol having another meaning can also be used as the character complement symbol.

According to the twelfth or twenty-eighth aspect of the present invention, the predicted character string candidates can be narrowed down to a character string desired by the user.

According to the invention of claim 13 or 29, the first
It is possible to increase the possibility of outputting the predicted character string desired by the user in the output priority order.

According to the invention of claim 14 or 30, the first
It is possible to further increase the possibility of outputting the predicted character string desired by the user in the output priority order.

According to the fifteenth or thirty-first aspect of the invention, it is possible to predict a character string following a hiragana, kanji, or hiragana character string mixed with kanji.

According to the sixteenth or thirty-second aspect of the present invention, the optimum character string prediction can be selected and performed according to the configuration of the input character string, and the flexibility of the apparatus can be improved.

According to the thirty-third aspect of the present invention, since it is possible to input the entire character string to be input into the document by inputting only the beginning portion of the character string to be input, the input efficiency of the character string is improved. Thus, the document can be created quickly and in a short time.

[Brief description of drawings]

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

2 is a flowchart showing a procedure for predicting a character string by the apparatus shown in FIG.

3 is a diagram specifically showing an example of predicting a character string by the device shown in FIG.

FIG. 4 is a block diagram showing another embodiment of the present invention.

5 is a flowchart showing a procedure for predicting a character string by the apparatus shown in FIG.

FIG. 6 is a block diagram showing still another embodiment of the present invention.

7 is a flowchart showing a procedure for predicting a character string by the apparatus shown in FIG.

8 is a diagram showing an example of contents of a learning table shown in FIG.

9 is a diagram showing an example of contents of the word dictionary shown in FIG.

10 is a diagram specifically showing an example of character string prediction by the apparatus shown in FIG.

FIG. 11 is a block diagram showing still another embodiment of the present invention.

12 is a flowchart showing a character string prediction operation procedure by the device shown in FIG.

13 is a diagram specifically showing an example of predicting a character string by the device shown in FIG.

FIG. 14 is a block diagram showing still another embodiment of the present invention.

15 is a flowchart showing a character string prediction operation procedure by the device shown in FIG.

FIG. 16 is a block diagram showing still another embodiment of the present invention.

FIG. 17 is a block diagram showing still another embodiment of the present invention.

[Explanation of symbols]

1 ... Input part 2 ... Input buffer 3 ... Dictionary search part 4 ... Word dictionary 5 ... Character string output part 6 ... Display part 7 ... Prediction candidate control part 8 ... Candidate buffer 9 ... Input learning part 10 ... Learning table 11 ... Flag buffer 12 ... Additional character number learning unit

Claims (33)

[Claims] 1. A character string following the input character string is predicted by searching a character string stored in the dictionary for a character string having the input character string at the head portion, and is searched from the dictionary. In a character string predicting method for predicting a character string following the input character string in a character string predicting device that outputs the generated character string as a predictive character string, an additional character number α is added to the character number n of the input character string ( (n + α) is an upper limit of the number of characters that the predicted character string has, and a predicted character string having a number of characters equal to or less than the upper limit value is searched from the dictionary. 2. The character string prediction method according to claim 1, wherein the added character number α is a variable having a functional relationship with the character number n of the input character string. 3. The character string prediction method according to claim 1, wherein the additional character number α is a preset fixed value. 4. Each time the predicted character string is output and is fixed by the user, the number of characters m of the fixed predicted character string is m.
And the number of input characters n at this time are collected, the relationship between the number of input characters n and the number of characters m is statistically processed, the additional character number α is obtained from the result of the statistical processing, and the obtained additional character number α is used from the dictionary. The character string prediction method according to claim 1, wherein the predicted character string is searched. 5. When there are a plurality of predicted character strings retrieved from the dictionary, the retrieved predicted character strings are rearranged in output priority order based on one or a plurality of criteria. Item 5. The character string prediction method according to any one of items 1 to 4. 6. The character string prediction method according to claim 5, wherein the output priority levels of these predicted character strings are rearranged on the basis of the number of characters of the predicted character strings retrieved from the dictionary. 7. The use frequency information of the character string stored in the dictionary is stored in this dictionary, and the output priority order of the predicted character string is based on the used frequency of the predicted character string retrieved from the dictionary. The character string prediction method according to claim 5, wherein the character strings are rearranged. 8. The input predictive character string is collected every time the predictive character string is output and fixed by the user, and when the character string is subsequently input by the user, the input character string is When a character string held at the beginning is searched from the dictionary and the collected definite predictive character strings, and as a result, there are a plurality of retrieved predictive character strings, from the collected definite predictive character strings. 8. The character string prediction method according to claim 5, wherein the output priority order of the predicted character strings is rearranged so that the output priority order of the searched predicted character strings is the highest priority and higher. 9. The collection of the definite predictive character string collects, in a learning table, an input character string that forms a head part of the definite predictive character string and a dictionary address of the dictionary that accommodates the definite predictive character string. 9. The character string prediction method according to claim 8, wherein the character string prediction method is performed by: 10. A character string following the input character string is predicted by searching a character string stored in the dictionary for a character string having the input character string at the head portion, and is searched from the dictionary. In a character string predicting method for predicting a character string following the input character string in a character string predicting device for outputting a character string as a predicted character string, when a user inputs a character completion symbol together with a character, it is input. Character string prediction having the same character at the same position as the character, and obtaining a predicted character string by searching the dictionary for a character string formed by complementing the character at the position of the input character completion symbol Method. 11. The predicted character string is not searched when the input character is only a character completion symbol, and the input character completion symbol is output as it is. String prediction method. 12. The character string according to claim 10, wherein a plurality of types of the character complement symbols are provided, and the type of the character to be complemented at this position is specified according to the type of the input character complement symbol. Prediction method. 13. When there are a plurality of retrieved predicted character strings, the retrieved predicted character strings are rearranged in output priority order based on one or a plurality of criteria. Any of the character string prediction methods described. 14. A confirmed predictive character string is collected every time the predictive character string is output and confirmed by a user,
When the user inputs a character string, the character string having this input character string at the beginning is searched from the dictionary and the collected definite predictive character string, and as a result, the searched predictive character string is If there are a plurality of predicted character strings, the output priority of these predicted character strings is rearranged so that the output priority of the predicted character strings retrieved from the collected predicted character strings is the highest priority. The character string prediction method according to claim 10. 15. The character string prediction method according to claim 1, wherein the character strings stored in the dictionary are hiragana, kanji, or hiragana mixed with kanji. 16. The method according to claim 1, wherein the input character string is a character string that does not include the character completion symbol.
The character string prediction method according to claim 10, wherein the predicted character string is obtained by any one of the character string prediction methods. 17. A character string following the input character string is predicted by searching the character string stored in the dictionary for a character string having the input character string at the beginning, and the character string is searched from the dictionary. In a character string predicting device for outputting the character string thus obtained as a predicted character string, a counting means for counting the number of characters in the character string, and a numerical value supplying means for supplying the additional character number α,
Addition means for adding the additional character number α supplied from the numerical value supply means to the character number n of the input character string counted by the counting means, and (n +
α) is the upper limit of the number of characters that the predicted character string has, and a search unit that searches the dictionary for a predicted character string having a number of characters equal to or less than the upper limit value, and the predicted character string searched by this search unit are displayed. A character string prediction device, comprising: a display unit for displaying the character string; and a selection unit for allowing a user to select an arbitrary character string as a confirmed predicted character from among the predicted character strings displayed by the display unit. 18. The character according to claim 17, wherein the added character number α supplied from the numerical value supply means is a variable having a functional relationship with the character number n of the input character string counted by the counting means. Column predictor. 19. The character string predicting apparatus according to claim 17, wherein the additional character number α supplied from the numerical value supply means is a fixed value preset in this means. 20. The number of characters m obtained by counting the number of characters of the confirmed predictive character string by the counting unit each time the predicting character string is confirmed by the selecting unit, and at this time, the number of characters already counted by the counting unit. Collecting means for collecting the input character number n, the calculating means for statistically processing the relationship of the character number m to the input character number n collected by the collecting means, and the added character number α from the statistical processing result by the calculating means. The addition value updating means for obtaining is provided, and the addition value supplying means is provided with the addition character number α obtained by the addition value updating means.
Is supplied to the search means.
The described character string prediction device. 21. When there are a plurality of predicted character strings retrieved from the dictionary by the retrieval means, the retrieved predicted character strings are rearranged in the output priority order based on one or a plurality of criteria, and then, 21. The character string prediction device according to claim 17, further comprising rearrangement means for outputting to the display means. 22. The rearranging means rearranges the output priority order of the retrieved predicted character strings on the basis of the number of characters counted by the counting means in the predicted character strings retrieved by the searching means. The character string prediction device according to claim 21. 23. A dictionary accommodating the usage frequency information of character strings is provided, and the rearranging means stores the predicted character strings searched based on the usage frequency of the predicted character strings searched from the dictionary by the searching means. 22. The character string prediction device according to claim 21, wherein the output priorities are rearranged. 24. A learning unit that collects a fixed predicted character string each time the predicting character string is fixed by the selecting unit; and a search unit when the character string is input by a user. When a character string having a string at the head portion is searched from the dictionary and the learning unit, and as a result, there are a plurality of predicted character strings searched, the rearrangement unit causes the predictive character searched from the learning unit. 24. The character string predicting device according to claim 21, wherein the character string predicting device performs rearrangement so that the columns are output with the highest priority with respect to other predicted character strings. 25. The learning means links a character string forming a head portion of a definite predictive character string and a dictionary address of the dictionary accommodating the definite predictive character string by collecting the character string, When the learning unit searches the learning unit for the same character string as the input character string, the searching unit searches the dictionary for a character string of the dictionary address collected by linking to the character string, and the searched character string is used for the learning. The predicted character string retrieved from the means is used.
4. The character string prediction device described in 4. 26. A character string following the input character string is predicted by searching the character string stored in the dictionary for a character string having the input character string at the head portion, and is searched from the dictionary. In a character string prediction device that outputs a generated character string as a predicted character string, an input means for inputting a character completion symbol together with a character, and an input character having the same character at the same position as the character input from this input means. Search means for searching the dictionary for a character string obtained by complementing characters at the position of the character completion symbol, display means for displaying the predicted character string searched by the search means, and display means for displaying the predicted character string. A character string prediction device, comprising: a selection unit that allows a user to select an arbitrary character string as a confirmed predicted character from among the predicted character strings. 27. When only the character completion symbol is input by the input means, the search means outputs the input character completion symbol as it is to the display means without searching for a predicted character string. 27. The character string prediction device according to claim 26. 28. A plurality of types of the character completion symbols input by the input means are used, and the search means complements a character of a specific type at the position of the character completion symbol according to the type of the character completion symbol. The character string prediction device according to claim 26 or 27. 29. When there are a plurality of predicted character strings retrieved from the dictionary by the retrieval means, the retrieved predicted character strings are rearranged in the output priority order based on one or a plurality of criteria, and then, 29. The character string predicting device according to claim 26, further comprising rearrangement means for outputting to a display means. 30. A learning unit that collects a fixed predicted character string each time the predicting character string is fixed by the selecting unit; and a search unit that inputs the character string when the user inputs the character string. When a character string having a string at the head portion is searched from the dictionary and the learning unit, and as a result, there are a plurality of predicted character strings searched, the rearrangement unit causes the predictive character searched from the learning unit. 30. The character string predicting device according to claim 29, wherein the character string predicting device is arranged to rearrange the strings so that the other predicted character strings are outputted with the highest priority. 31. The character string prediction device according to claim 17, wherein the character strings stored in the dictionary are hiragana, kanji, or hiragana character strings mixed with kanji. 32. A determining unit is provided for determining whether the character string input by the input unit is a character string that does not include the character completion symbol, and the input character string is the character completion by the determining unit. 27. The character string prediction device according to claim 26, wherein when it is determined that the character string does not include a symbol, the predicted character string is obtained by the character string prediction device according to any one of claims 17 to 25. 33. A document creating apparatus for creating a document by inputting a character string, wherein the character string predicting apparatus according to claim 17 is mounted, and a predictive character string determined by the character string predicting apparatus is provided. A document creating apparatus characterized by inputting in the document.