JPH08255159A - Character string prediction method, character string prediction device, kana-kanji conversion method, kana-kanji conversion device, and document creation device - Google Patents

Abstract

(57) [Summary] [Purpose] When predicting a character string composed of Kanji from an input character string of Hiragana, the output priority of the predicted character string desired by the user should be set to a higher priority. [Configuration] The Kanji conversion unit 2 converts the Hiragana character string input from the input unit 1 into Kanji by referring to the dictionary 3, and obtains conversion candidates. The stroke number checking unit 4 checks the stroke number of the conversion candidate with reference to the dictionary 5, adds the stroke number to the conversion candidate, and passes it to the character string rearranging unit 6. The character string rearranging unit 6 rearranges the conversion candidates in the order of the number of strokes, and passes the conversion candidates to the predicted character string searching unit 7 in that order. The predictive character string search unit 7 searches the dictionary 8 for predictive character string candidates of the received conversion candidates in the order of reception,
The retrieved predicted character string candidates are passed to the character string output unit 9 in the search order. The character string output unit 9 displays the predicted character string candidates in the order passed to the display unit 10, and if the displayed predicted character string candidate is the desired one, the user selects it and if not. , Display the predicted character string of the next candidate.

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 following a hiragana character string input by an input device capable of inputting kanji and kana, a character string prediction device and a handwriting using this method. The present invention relates to a kana-kanji conversion method for converting an input character string into kanji, a kana-kanji conversion device using this method, and a document creation device equipped with the character string prediction device or the kana-kanji conversion device.

[0002]

2. Description of the Related Art Conventionally, when a character is input from an input device such as a document creation device, it takes time and effort to input a character string, especially when inputting a character string using a handwritten character recognition device or the like. Unlike the input from the keyboard, it takes time to input one character, which tends to increase the burden on the user. Therefore, in order to reduce the burden on the user, when a part of the characters (strings) to be entered is entered, the character string that follows the characters (strings) that have been entered up to that point is predicted, and the user A method of inputting the entire character string intended for input has been considered. However, depending on the input character string, there may be many predicted character strings. In such a case, if the user does not take priority to output the character string desired by the user, Had a problem that it took time to determine the predicted character string.

Therefore, in order to narrow down the predicted character string desired by the user, the previously used words are learned, the usage frequency information of the character string is used, the address is input, the name is input, and the like. Various methods such as limiting By the way, when predicting a character string that follows an input character string composed of Kanji, many Kanji have multiple readings, so simply predict the character string that follows from the character string of the input Kanji. Then, the number of predicted character strings will increase, and therefore the output priority of the predicted character strings must be set. However, if you do this easily, there is a possibility that the priority of the character string that is different from the reading intended by the user will be higher, and there is no way to effectively assign the output priority to the predicted character string. There was a problem.

Further, when the user inputs a part of the beginning of a character string composed of kanji characters in hiragana, the information for determining the kind of kanji character forming the character string following the input character string is left as it is. Since the above problem is not included at all, the problem described above is further promoted, and there is a problem that the prediction accuracy of the character string deteriorates.

Further, in the case of inputting characters in hiragana from a handwritten character input device to convert into kana-kanji, the user tends to input kanji with a large number of strokes in a hiragana character string to convert into kana-kanji. Since a kana-kanji conversion method considering such a tendency has not been developed, there is a problem that the kana-kanji conversion efficiency is poor.

[0006]

In the above-described character string prediction method, when the input character string is a hiragana character and the character string composed of the kanji following this input character string is predicted, the above-mentioned hiragana input is performed. Since the character string does not include any information for predicting a character string composed of Kanji, it is not possible to determine the priority of the predicted character string,
There is a problem that the accuracy of predicting the character string deteriorates, and it takes time for the user to determine the predictive character string.
In addition, when converting a character string input in hiragana from a handwritten character input device into kana-kanji, the kana with a larger number of strokes does not consider the tendency for the user to enter in hiragana, so the kana-kanji conversion efficiency is poor. was there.

Therefore, the present invention has been made in consideration of the above circumstances, and a predicted character string candidate obtained without impairing the prediction accuracy of a predicted character string composed of Kanji characters based on a Hiragana input character string. A character string prediction method capable of improving the operability by limiting the options of the character string prediction apparatus, a character string prediction apparatus using the method, and a document creation apparatus equipped with the character string prediction apparatus, and a handwriting input apparatus. Provide a kana-kanji conversion method that can improve the conversion efficiency when the hiragana character string input in step 2 is converted to kana-kanji, a kana-kanji conversion device using this method, and a document creation device equipped with this kana-kanji conversion device. Is intended.

[0008]

According to a first aspect of the present invention, a character string having the input character string at its head portion is searched from the character strings stored in the dictionary, and the searched character string is predicted. In a character string prediction method for predicting a character string following the input character string in a character string prediction device for outputting as a character string, it is determined whether the input character string is a Hiragana character string, and the character string is a Hiragana character string. In this case, when the input character string is converted into Kanji and a Kanji conversion candidate is obtained, the number of strokes of each Kanji constituting the Kanji conversion candidate is obtained, and thereafter, the number of strokes of each Kanji constituting one Kanji conversion candidate is totaled. If the number of strokes of each Kanji conversion candidate is multiple, the output priority of the predicted character string that has these Kanji conversion candidates at the head is determined, and then these Kanji characters are used. Conversion The character string having the head portion is retrieved from the string contained in the dictionary, and the predicted character string obtained by adopting the method of outputting the output priority order.

The invention of claim 2 employs a method in which the output priority of the predicted character string is set higher as the number of strokes of each Kanji conversion candidate increases.

According to the invention of claim 3, the kanji conversion candidates are rearranged in the output priority order, the predicted character strings of the respective kanji conversion candidates are searched from the dictionary in the rearranged order, and the predicted characters are searched in the searched order. The method of outputting columns is adopted.

According to a fourth aspect of the present invention, a character string having the input character string at the beginning is searched from the character strings stored in the dictionary, and the searched character string is output as a predicted character string. In a character string prediction method for predicting a character string following the input character string in a column prediction device, it is determined whether the input character string is a Hiragana character string, and if it is a Hiragana character string, the input character When a column is converted to Kanji and a Kanji conversion candidate is obtained, the number of strokes of each Kanji that constitutes the Kanji conversion candidate is calculated, and then the number of strokes of each Kanji that constitutes one Kanji conversion candidate is summed up. The number of strokes, and if there are multiple Kanji conversion candidates, the output priority of these Kanji conversion candidates is determined by the number of strokes of each Kanji conversion candidate, and then these Kanji conversion candidates are displayed in this output priority order and the user is displayed. Waiting for selection, then searching for a character string that has the Kanji conversion candidate selected by the user at the beginning from the character strings stored in the dictionary, and outputting the obtained predicted character string. are doing.

According to the invention of claim 5, the kanji conversion candidate is 1
When the character is composed of Chinese characters, a method is adopted in which the stroke number of this Chinese character is used as the stroke number of the Kanji conversion candidate.

According to a sixth aspect of the present invention, when the input character string is not a hiragana character string, it is further determined whether or not the input character string is a kanji character. When the input character string is a kanji character, the input character string of the kanji character is saved. After that, the same Kanji candidate as the stored Kanji character string is removed from the Kanji candidates obtained by converting the input character string of Hiragana, and the output priority is determined for the remaining Kanji candidates. It employs the method of performing the subsequent processing.

According to a seventh aspect of the present invention, when the input character string is not a Hiragana character string, a predictive character string having this input character string at the beginning is immediately retrieved from the dictionary and output. .

According to an eighth aspect of the invention, when there are a plurality of predicted character strings having the same Kanji conversion candidate or the same input character string at the head portion, the output priority order is determined based on a predetermined criterion, The method of outputting according to the determined output priority order is adopted only between predicted character strings.

The invention of claim 9 employs a method of setting the number of the predicted character strings to be searched to be a preset number or less.

According to a tenth aspect of the present invention, a character string that has the input character string at the beginning is searched from the character strings stored in the dictionary, and the searched character string is output as a predicted character string. In the column prediction device, a determining unit that determines whether the input character string is a Hiragana character string, and if the determining unit determines that the input character string is a Hiragana character string, the input character string is a Kanji character. Kana-kanji conversion means for converting to kanji to obtain kanji conversion candidates, stroke number acquisition means for obtaining the stroke number of each kanji constituting the kanji conversion candidate obtained by this kana-kanji conversion means, and one kanji conversion candidate The stroke number calculation means for totaling the stroke numbers obtained by the stroke number acquisition means for each kanji as the stroke number of this kanji conversion candidate, and the stroke number calculation means when there are a plurality of kanji conversion candidates Comparing means for comparing the number of strokes of each kanji conversion candidate, deciding means for deciding the output priority order of the predicted character string having the kanji conversion candidate at the head based on the comparison result of this comparison means, and the kanji conversion candidate. Retrieval means for retrieving the character string held at the head part from the character strings contained in the dictionary, and a display for displaying the predictive character string retrieved by this retrieving means in the order of the output priorities determined by the deciding means. And a confirmation unit that allows the user to select an arbitrary character string as a confirmed predicted character string from among the predicted character strings displayed by the display unit.

According to an eleventh aspect of the present invention, the determining means determines that the output priority of the predicted character string becomes higher as the number of strokes of each Kanji conversion candidate increases.

According to a twelfth aspect of the present invention, there is provided sorting means for sorting the kanji conversion candidates in the output priority order determined by the determining means, and the searching means is arranged in the order sorted by the sorting means. The prediction candidate character strings of conversion candidates are searched from the dictionary, and the output means outputs the predicted character strings in the order searched by the searching means.

According to a thirteenth aspect of the present invention, a character string that has the input character string at the beginning is searched from the character strings stored in the dictionary, and the searched character string is output as a predicted character string. In the column prediction device, a determining unit that determines whether the input character string is a Hiragana character string, and if the determining unit determines that the input character string is a Hiragana character string, the input character string is a Kanji character. Kana-kanji conversion means for converting to kanji to obtain kanji conversion candidates, stroke number acquisition means for obtaining the stroke number of each kanji constituting the kanji conversion candidate obtained by this kana-kanji conversion means, and one kanji conversion candidate The stroke number calculation means for totaling the stroke numbers obtained by the stroke number acquisition means for each kanji as the stroke number of this kanji conversion candidate, and the stroke number calculation means when there are a plurality of kanji conversion candidates Comparing means for comparing the number of strokes of each kanji conversion candidate, deciding means for deciding the output priority of the predicted character string having the kanji conversion candidate at the head based on the comparison result of this comparing means, and deciding by this deciding means. Display means for displaying the kanji conversion candidates obtained by the kana-kanji conversion means in the order of the output priority, selection means for allowing the user to select any kanji conversion candidate displayed by the display means, and this selection means. Searching means for searching a character string having the Kanji conversion candidate selected by the head part from the character strings stored in the dictionary, and displaying means for displaying the predicted character string obtained by the searching means. The predictive character string displayed by the display means is provided with a confirming means for allowing the user to select an arbitrary character string as the confirmed predictive character string.

According to a fourteenth aspect of the present invention, the stroke number calculating means has a configuration in which, when the Kanji conversion candidate is composed of one Kanji character, the stroke number of this Kanji character is taken as the stroke number of the Kanji conversion candidate.

According to a fifteenth aspect of the present invention, when the determining unit determines that the input character string is not a Hiragana character string,
Further, a sub-determination means for determining whether or not this input character string is a kanji character is provided, and when the input character string is determined to be a kanji character by this sub-determination means, a storage means for saving the input character string of this kanji character. An identifying unit for identifying the same Kanji character string as the Kanji character string stored in the storing unit from the Kanji candidate obtained by the Kana-Kanji converting unit; and the Kanji character string identified by the identifying unit, The removing means for removing the kanji candidates from the kanji candidates, and the deciding means are configured to decide the output priority order for the remaining kanji candidates not removed by the removing means.

According to the sixteenth aspect of the present invention, when the determining unit determines that the input character string is not a Hiragana character string,
The search means immediately searches the dictionary for a predicted character string having the input character string at the head portion thereof, and the output means outputs the searched predicted character string.

According to the seventeenth aspect of the invention, when there is a plurality of predicted character strings having the same Kanji conversion candidate or the same input character string at the beginning as a result of the search by the search means, output of these predicted character strings has priority. Sub-determination means for determining the order based on a predetermined standard is provided, and the output means is configured to output the predictive character string whose output priority order is determined by the sub-determination means in the output priority order. .

According to the eighteenth aspect of the present invention, the search means has a configuration in which the number of predicted character strings searched from the dictionary is set within a preset number.

The invention of claim 19 is a document creating apparatus for creating a document by predicting a character string.
The character string prediction device according to any one of 8 is installed, and the predicted character string determined by the character string prediction device is input into the document.

The invention of claim 20 is a kana-kanji conversion method in a kana-kanji conversion device for converting an input character string into kana-kanji, wherein the number of strokes of each kanji constituting a kanji candidate obtained by kana-kanji conversion. Then, the number of strokes of each kanji composing one kanji candidate is summed up to the number of strokes of this kanji candidate, and if there are multiple kanji candidates, the output of these kanji conversion candidates is determined by the number of strokes of each kanji candidate. After determining the priorities, a method of displaying these Kanji conversion candidates in the order of the output priorities is adopted.

According to a twenty-first aspect of the present invention, in a kana-kanji conversion device for converting an input character string into kana-kanji characters, kana-kanji conversion means for converting the input character string into kanji to obtain kanji conversion candidates, and the kana-kanji conversion character. A stroke number obtaining means for obtaining the stroke number of each kanji constituting the kanji conversion candidate obtained by the converting means;
A stroke number calculating means for summing up the stroke numbers obtained by the stroke number acquiring means of each kanji constituting one kanji conversion candidate to obtain the stroke number of this kanji conversion candidate, and the stroke number calculation when there are a plurality of kanji conversion candidates. Comparing means for comparing the number of strokes of each kanji conversion candidate obtained by the means, and deciding means for deciding the output priority of the predicted character string having the kanji conversion candidate at the head according to the comparison result of the comparing means; And a display unit for displaying the Kanji conversion candidates obtained by the Kana-Kanji conversion unit in the order of output priority determined by the determination unit.

According to a twenty-second aspect of the present invention, there is provided a document creating apparatus comprising handwriting input means for inputting characters by handwriting, wherein the character string input by the handwriting input means is converted into kana and kana and kanji and input. The kana-kanji conversion device according to item 21 is installed, and when the input character string is hiragana, the kana-kanji conversion device is used when converting the hiragana character string into kana-kanji characters.

[0030]

According to the character string prediction method of the invention of claim 1,
It is determined whether the input character string is a Hiragana character string, and if the character string is a Hiragana character string, the input character string is converted to Kanji to obtain a Kanji conversion candidate. The number of strokes is calculated, and then the number of strokes of each Kanji character that constitutes one Kanji conversion candidate is summed up to be the stroke number of this Kanji conversion candidate. If there are multiple Kanji conversion candidates, these are determined by the number of strokes of each Kanji conversion candidate. After determining the output priority of the predicted character string having the Kanji conversion candidate at the beginning, the character string having these Kanji conversion candidates at the beginning is searched from the character strings stored in the dictionary, and the obtained prediction is obtained. The character strings are output in the output priority order.

In the character string prediction method according to the second aspect of the present invention, the higher the number of strokes of each Kanji conversion candidate is, the higher the output priority of the predicted character string is.

In the character string prediction method of the invention of claim 3, the kanji conversion candidates are rearranged in the output priority order, and the predicted character strings of the respective kanji conversion candidates are retrieved from the dictionary in the rearranged order. , The predicted character strings are output in the searched order.

According to the character string prediction method of the invention of claim 4, it is judged whether or not the input character string is a hiragana character string, and if it is a hiragana character string, the input character string is converted into kanji. When the kanji conversion candidates are obtained, the number of strokes of each kanji constituting the kanji conversion candidate is calculated, and then the number of strokes of each kanji constituting one kanji conversion candidate is summed to be the number of strokes of the kanji conversion candidate, and the kanji When there are multiple conversion candidates, after determining the output priority of these Kanji conversion candidates according to the number of strokes of each Kanji conversion candidate, these Kanji conversion candidates are displayed in this output priority order and wait for the user's selection, After that, the character string having the Kanji conversion candidate selected by the user at the head portion is searched from the character strings stored in the dictionary, and the obtained predicted character string is output.

In the character string prediction method of the present invention, when the Kanji conversion candidate is composed of one Kanji character, the stroke number of this Kanji character is taken as the stroke number of the Kanji conversion candidate.

According to the character string prediction method of the invention of claim 6, when the input character string is not a hiragana character string, it is further judged whether or not the input character string is a kanji character. The input character string of the kanji is saved, and thereafter, the same kanji candidate as the saved kanji character string is removed from the kanji candidates obtained by converting the input character string of hiragana, and the remaining kanji candidates are removed. The process after determining the output priority order is performed.

According to the character string prediction method of the invention of claim 7, when the input character string is not a hiragana character string, a predicted character string having this input character string at the head portion is immediately retrieved from the dictionary, Output.

According to the character string prediction method of the invention of claim 8, when there are a plurality of predicted character strings having the same Kanji conversion candidate or the same input character string at the head portion, based on a predetermined criterion. The output priority is determined, and only the predicted character strings are output according to the determined output priority.

In the character string prediction method of the ninth aspect of the present invention, the predicted character strings of a preset number or less are searched.

In the character string predicting apparatus of the tenth aspect of the invention, the judging means judges whether or not the input character string is a Hiragana character string. When the determining unit determines that the input character string is a Hiragana character string, the Kana-Kanji converting unit converts the input character string into Kanji and obtains a Kanji conversion candidate. The stroke number obtaining means obtains the stroke number of each kanji constituting the kanji conversion candidate obtained by the kana-kanji conversion means. The stroke number calculation means sums up the stroke numbers of the respective kanji characters forming one kanji conversion candidate obtained by the stroke number acquisition means to obtain the stroke number of this kanji conversion candidate. The comparing means compares the number of strokes of each kanji conversion candidate obtained by the stroke number calculating means when there are a plurality of kanji conversion candidates. The deciding means decides the output priority order of the predicted character string having the Kanji conversion candidate at the head according to the comparison result of the comparing means. The search means searches the character string stored in the dictionary for a character string having the Kanji conversion candidate at the beginning. The display means displays the predicted character strings retrieved by the retrieval means in the output priority order determined by the determination means.
The confirming means allows the user to select an arbitrary character string as the confirmed predictive character string from among the predictive character strings displayed by the display means.

In the character string prediction device according to the invention of claim 11, the deciding means has a greater number of strokes of each Kanji conversion candidate,
It is determined that the output priority of the predicted character string is high.

In the character string prediction apparatus of the twelfth aspect of the present invention, the sorting means sorts the Kanji conversion candidates in the output priority order determined by the determining means. The searching means searches the dictionary for predicted character strings of each Kanji conversion candidate in the order sorted by the sorting means, and the output means outputs the predicted character strings in the order searched by the searching means. .

In the character string predicting device of the thirteenth aspect of the present invention, the judging means judges whether or not the input character string is a hiragana character string. When the determining unit determines that the input character string is a Hiragana character string, the Kana-Kanji converting unit converts the input character string into Kanji and obtains a Kanji conversion candidate. The stroke number obtaining means obtains the stroke number of each kanji constituting the kanji conversion candidate obtained by the kana-kanji conversion means. The stroke number calculation means sums up the stroke numbers of the respective kanji characters forming one kanji conversion candidate obtained by the stroke number acquisition means to obtain the stroke number of this kanji conversion candidate. The comparing means compares the number of strokes of each kanji conversion candidate obtained by the stroke number calculating means when there are a plurality of kanji conversion candidates. The deciding means decides the output priority order of the predicted character string having the Kanji conversion candidate at the head according to the comparison result of the comparing means. The display means displays the kanji conversion candidates obtained by the kana-kanji conversion means in the output priority order determined by the determination means. The selection unit causes the user to select an arbitrary Kanji conversion candidate displayed by the display unit. The searching means searches the character strings stored in the dictionary for a character string having the Kanji conversion candidate selected by the selecting means at its head portion. The display means displays the predicted character string obtained by the search means. The confirming means allows the user to select an arbitrary character string as the confirmed predictive character string from among the predictive character strings displayed by the display means.

In the character string prediction device of the fourteenth aspect of the invention, when the kanji conversion candidate is composed of one kanji character, the stroke number calculating means sets the stroke number of this kanji character as the stroke number of the kanji conversion candidate.

In the character string predicting device of the fifteenth aspect of the present invention, when the sub-judging means judges that the input character string is not a hiragana character string by the judging means, it is further determined whether or not the input character string is a kanji character. judge. When the sub determination unit determines that the input character string is a Chinese character, the storage unit stores the input character string of the Chinese character. The identifying means identifies, from the kanji candidates obtained by the kana-kanji conversion means, a kanji candidate that is the same as the kanji character string stored in the storage means. The removing unit removes the Kanji character string identified by the identifying unit from the Kanji candidates. The determining unit determines the output priority order for the remaining Kanji candidates that have not been removed by the removing unit.

In the character string predicting apparatus of the sixteenth aspect of the invention, when the determining unit determines that the input character string is not a Hiragana character string, the searching unit determines the predictive character string having the input character string at the head portion. Is immediately retrieved from the dictionary, and the output means outputs the retrieved predicted character string.

In the character string predicting device of the seventeenth aspect of the invention, the sub determining means, as a result of the search by the searching means, has a plurality of predicted character strings having the same Kanji conversion candidate or the same input character string at the head portion. In this case, the output priority order of these predicted character strings is determined based on a predetermined standard. The output means outputs the predicted character strings whose output priority order is determined by the sub-determination means in this output priority order.

In the character string predicting device of the eighteenth aspect of the present invention, the searching means searches the dictionary for a predetermined number of predicted character strings or less.

In the document creating apparatus of the nineteenth aspect of the invention, the predicted character string determined by the character string predicting apparatus of any one of the tenth to eighteenth aspects is input into the document.

In the kana-kanji conversion method of the invention of claim 20, the number of strokes of each kanji forming a kanji candidate obtained by kana-kanji conversion is obtained, and then each kanji forming one kanji candidate is calculated. The total number of strokes is taken as the stroke count of this Kanji candidate, and when there are multiple Kanji candidates, after determining the output priority of these Kanji conversion candidates by the number of strokes of each Kanji candidate,
These Kanji conversion candidates are displayed in the output priority order.

In the kana-kanji conversion device of the twenty-first aspect of the invention, the kana-kanji conversion means converts the input character string into kanji to obtain kanji conversion candidates. The stroke number obtaining means obtains the stroke number of each kanji constituting the kanji conversion candidate obtained by the kana-kanji conversion means. The stroke number calculation means sums up the stroke numbers of the respective kanji characters forming one kanji conversion candidate obtained by the stroke number acquisition means to obtain the stroke number of this kanji conversion candidate. The comparing means compares the number of strokes of each kanji conversion candidate obtained by the stroke number calculating means when there are a plurality of kanji conversion candidates. The deciding means decides the output priority order of the predicted character string having the Kanji conversion candidate at the head according to the comparison result of the comparing means. The display means displays the kanji conversion candidates obtained by the kana-kanji conversion means in the output priority order determined by the determination means.

In the document creating apparatus of the twenty-second aspect of the invention, the kana-kanji conversion apparatus of the twenty-first aspect converts the hiragana character string into kana-kanji characters when the input character string is hiragana.

[0052]

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, a Kanji conversion unit 2, a kana character string / Kanji dictionary 3, a stroke number investigation unit 4, a Kanji / stroke number dictionary 5, a character string rearrangement unit 6, a predicted character string search unit 7, and a word dictionary. 8, a character string output unit 9 and a display unit 10.

The function of each block described above will now be described. The input unit 1 comprises a handwritten character recognition device,
Characters (strings) obtained by recognizing handwritten characters (strings) are input. The kanji conversion unit 2 converts the input character string into a kanji character string by referring to the kana character string / kanji dictionary 3 (kana kanji conversion). As shown in FIG. 3, the kana character string / kanji dictionary 3 contains kanji whose reading is a hiragana character string. The stroke number investigation unit 4 searches the number of strokes of each kanji constituting the kanji conversion candidate obtained by converting the kana to kanji by referring to the kanji / stroke number dictionary 5, and determines the number of strokes of the kanji constituting each kanji conversion candidate. Give the grand total. However, when the kanji conversion candidate is composed of one kanji, the number of strokes of this kanji is taken as the number of strokes of the kanji conversion candidate. As shown in FIG. 4, the kanji / stroke count dictionary 5 stores the stroke count for each kanji. Character string sorting unit 6
Rearranges the output priority of Kanji conversion candidates according to the number of strokes. The character string search unit 7 searches for a predicted character string candidate by searching the word dictionary 8 based on the passed Kanji conversion candidate (input character string). As shown in FIG. 5, the word dictionary 8 is registered with words for searching for predicted character string candidates using a character string of one or more characters as a key. The character string output unit 9 outputs the passed predicted character string candidates to the display unit 10. The display unit 10 includes an LCD, a CRT, etc., and displays an input character string, a predicted character string candidate, and the like.

Next, the operation of this embodiment will be described.
In this example, description will be given only when predicting a character string composed of Kanji. 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 an internal buffer. When the prediction key is pressed, the input unit 1 sets the character string held in the buffer until then as an input character string, It is determined whether this input character string is a Hiragana character string. As a result, when the input character string is a Hiragana character string, the input unit 1 passes the input character string to the Kanji conversion unit 2, and when the input character string is a Kanji character, the input character string (Kanji) is directly input to the predictive character string search unit 7. Pass to.

For example, if the prediction key is pressed when the user inputs "to", "to" is output from the internal buffer of the input unit 1 to the Kanji conversion unit 2 as an input character string and used. If the prediction key is pressed when the person inputs "tori", "tori" is output from the buffer to the kanji conversion unit 2 as an input character string. However, if the user presses the prediction key at the time when the user inputs, for example, "Tou", this "Tou" is directly output from the buffer to the predicted character string search unit 7 as an input character string.

The Kanji conversion unit 2 takes out the Kanji for reading the input Hiragana character string from the Kana character string / Kanji dictionary 3, and passes the taken out Kanji to the stroke number investigation unit 4. For example, if the Hiragana character string "To" is entered,
The kanji conversion unit 2 takes out kanji such as "ten", "east", "throw" from the kana character string / kanji dictionary 3 as shown in FIG.
This is output to the stroke number investigation unit 4 as a conversion candidate. The stroke number investigation unit 4 retrieves the stroke number of each kanji composing the passed kanji conversion candidate by referring to the kanji / stroke number dictionary 5 as shown in FIG. 4, and obtains the stroke number of each kanji constituting one kanji conversion candidate. The number of strokes obtained by summing up all of them is set as the number of strokes of the kanji conversion candidate.

The number of strokes of each Kanji conversion candidate obtained in this way is output to the character string rearrangement unit 6 as a pair with the conversion candidate as shown in FIG. However, when the kanji conversion candidate is composed of one kanji, the stroke number of this kanji becomes the stroke number of the kanji conversion candidate. The character string rearrangement unit 6 rearranges the input conversion candidates such that the larger the number of strokes, the higher the output priority, and outputs the conversion candidates to the predicted character string search unit 7 in this rearranged order. To do.

The predictive character string retrieval unit 7 is a character string rearranging unit 6
When the input character strings converted from Kanji to Kanji are received in order,
With respect to the received input character string, the word dictionary 8 as shown in FIG. 5 is referred to search for a predicted character string including the input character string at the head. The search order at this time is the order of the input character strings received by the predicted character string search unit 7, that is, the output priority order described above. The predictive character string search unit 7 passes the predictive character string candidates obtained by the search to the predictive character string output unit 9 in the search end order.

On the other hand, when the predictive character string retrieving unit 7 directly receives the input character string of the Kanji from the input unit 1, the predictive character string having these Kanji conversion candidates at the head portion is retrieved from the word dictionary 8 and the obtained predictive character string is obtained. The character string candidates are passed to the character string output unit 9.
In this case, the output order of the obtained predicted character string candidates to the character string output unit 9 is not prioritized. However, since the input character string is Kanji from the beginning and the number of character strings to be predicted is one, the number of predicted character string candidates to be searched is smaller than in the case of Hiragana character string input, so prioritization is performed. Even if it is not done, it will not cause much trouble.

The character string output unit 9 displays the predicted character string candidates passed from the predicted character string search unit 7 on the display unit 10 in the order in which they are received (in the output priority order). It should be noted that the character string output unit 9 displays the passed predicted character string candidates in the order in which they are passed, even if they are not prioritized.

Here, the concept of the above-mentioned output prioritization will be described. When a user wants to input a character string of Kanji and wants to predict a part of the character string that follows by using Hiragana, the more Kanji with more strokes, the more time it takes to input it in Kanji. Tends to type in. Therefore, it is considered that the higher the number of strokes in the conversion candidates obtained by converting the input hiragana into the kana-kanji character, the higher the possibility of forming the head portion of the character string that the user wants to input. Therefore, the predicted character string candidates are searched in the order of the number of strokes of the conversion candidates, and the searched predicted character string candidates are output to the display unit 10 in the searched order, so that the output prioritization of the above-described concept is performed. Will be done.

FIG. 2 is a flow chart showing the procedure of the character prediction process described above. The input unit 1 inputs a character string in step 201, and when the character string becomes a prediction target, it determines whether the character string input in step 202 is a hiragana character string. As a result, if the character string is not a Hiragana character string, the input unit 1 proceeds to step 208 to pass the input character string to the predictive character string retrieval unit 7, and if it is a Hiragana character string, proceeds to step 203 to convert the input character string into Kanji characters. Deliver to Part 2. The kanji conversion unit 2 converts the hiragana character string received in step 203 into a kanji character string by referring to the kana character string / kanji dictionary 3, and passes the kanji conversion candidates obtained by the conversion to the stroke number investigation unit 4.

The stroke number investigation unit 4 determines the stroke number of each Kanji character forming the Kanji conversion candidate received in step 204.
The stroke number dictionary 5 is searched, the stroke number is calculated for each Kanji conversion candidate, and the obtained stroke number is passed to the character string rearranging unit 6 together with the Kanji conversion candidate. The character string rearranging unit 6 is step 205.
After rearranging the kanji conversion candidates in the order of the number of strokes of the kanji conversion candidates (corresponding to prioritization), the kanji conversion candidates are output to the predicted character string search unit 7 in the rearranged order.

The predictive character string search unit 7 searches the word dictionary 8 for a corresponding predictive character string candidate for each of the Kanji conversion candidates received in step 206, and obtains the predictive character string candidates in the order in which they are searched. It is passed to the output unit 9. However, for the kanji character string directly passed from the input unit 1, the corresponding predicted character string candidate is searched from the word dictionary 8 and passed to the character string output unit 9.

The predicted character string output unit 9 displays the predicted character string candidates on the display unit 10 in the passed order (priority order) in step 207. However, although the predicted character string candidates of the input character string directly input in the Kanji from the input unit 1 are not prioritized, they are displayed in the same order. When the predicted character string candidate displayed on the display unit 10 is the desired one, the user presses the select / execute key of the input unit 1 to select it, thereby fixing the predicted character string. When the predicted character string candidate displayed on the display unit 10 is not the desired one, the user presses the next candidate key on the input unit 1 to display the predicted character string candidate of the next output priority on the character string output unit 9. It is displayed on the display unit 10. The character string output unit 9 may display the passed predicted character string candidates in a list on the display unit 10, and the user may select and confirm the desired predicted character string.

Next, the operation of the above embodiment will be described using a specific example. When the hiragana character string “TO” is input from the input unit 1, the Kanji conversion unit 2 receives the character string “TO”. The kanji conversion unit 2 retrieves a conversion candidate for the hiragana character string "to" from the kana character string / kanji dictionary 3 as shown in FIG. 3 and obtains "ten, east, throw". Since these Kanji conversion candidates are passed to the stroke count research unit 4, the stroke count research unit 4 retrieves and receives the stroke count for each character of "10, east, throw" from the Kanji / stroke count dictionary 5 shown in FIG. Correct the output format such as "Ten (2), East (8), Throw (7)" so that the Kanji and the stroke number of the Kanji are paired, and pass them to the character string rearrangement unit 6. . The character string rearranging unit 6 rearranges the conversion candidates received in descending order of the number of strokes, here the number of strokes in parentheses, to obtain "east (8), throw (7), and ten (2)". The order at this time becomes the output priority of the kanji conversion candidates as it is.

The predictive character string search unit 7 receives the kanji conversion candidates in the order in which they are received, that is, in descending order of priority of the kanji conversion candidates that have been prioritized, the words having the conversion candidates at the beginning, that is, the predicted characters. A column candidate is searched from the word dictionary 8 shown in FIG. When the search is completed for all the candidates, they are passed to the character string output unit 9 in the order in which the search is completed. In this example, the predicted character string candidates “Tokyo, Tokai, pitcher, pitcher, ten” can be obtained in this order, and they are arranged in the search end order (output priority order). Therefore, the character string output unit 9 receives these predicted character string candidates in the order of "Tokyo, Tokai, pitcher, pitch, ten", and then displays them on the display unit 10 in the received order. That is, the character string output unit 9 displays "Tokyo" in the first rank, and when the next candidate key of the input unit 1 is pressed, "Tokai" is displayed in the second rank. The same applies thereafter.
At this time, the character string output unit 9 confirms and displays the displayed predicted character string when the selection / execution key for selecting the predicted character string candidate displayed on the display unit 10 is pressed from the input unit 1. By the way, when the kanji character string “east” is input to the input unit 1 from the beginning, the kanji character string “east” is directly passed to the predicted character string detection unit 7. As a result, the predicted character string detection unit 7 searches the word dictionary 8 for a predicted character string candidate “Tokyo, Tokai” having “East” at the head, and the character string output unit 9
Pass to.

The character string output unit 9 displays "Tokyo, Tokai" on the display unit 10 in the order in which they are passed. In this case, the input character string that is the key when searching the predicted character string in the predicted character string detection unit 7 is "East", so the number of predicted character string candidates to be searched is the number of input hiragana characters. Compared to the case (2/5 in this example), the number of predictive character string candidates displayed on the display unit 10 is small even without prioritization, so that the selection at the time of confirmation is not complicated. Absent.

According to the present embodiment, when predicting a character string composed of kanji and the character string is input in hiragana, the hiragana input character string is converted into kanji, and at this time, a plurality of conversions are performed. When a candidate is obtained, the number of strokes of each conversion candidate is checked, rearranged in the order of the number of strokes, and a predicted character string candidate having each conversion candidate at the head in this order is searched, and the predicted character string obtained at this time By setting the output priority order of the candidates as the search order, it is possible to increase the output priority order of the predicted character string candidates intended for input by the user and improve the selection efficiency when determining the predicted character string. It should be noted that the priority order between predicted character string candidates obtained from the same conversion candidate can be easily assigned by storing the usage frequency of each predicted character string in the word dictionary 8 or the like and using this usage frequency. It is possible to further improve the selection efficiency of the predicted character string.

Further, in the present example, description has been made on the premise of predicting a character string composed of Kanji characters, but it is easy to use it together with prediction of a Hiragana character string or a Hiragana character string mixed with Kanji. When the candidates are composed of kanji, the same effect can be obtained by prioritizing this example.

FIG. 6 is a block diagram showing another embodiment of the present invention. The character string prediction device includes an input unit 1, a Kanji conversion unit 2, a kana character string / Kanji dictionary 3, a stroke number investigation unit 4, a Kanji / stroke number dictionary 5, a character string rearrangement unit 6, a predicted character string search unit 7,
It comprises a word dictionary 8, a character string output unit 9, a display unit 10, a used Kanji removal unit 11, and a used Kanji buffer 12.

The functions of blocks not used in the embodiment of FIG. 1 among the above blocks will be described below. The used kanji removal section 11 removes the kanji registered in the used kanji buffer 12 from the conversion candidates converted into kanji by the kanji conversion section 2 and passes it to the stroke number investigation section 4. The used kanji character buffer 12 stores the kanji input from the input unit 1 from the beginning as shown in FIG.

Next, the operation of this embodiment will be described with reference to the flowchart of FIG. The input unit 1 is step 70
When a character string is input in 1 and becomes a prediction target, it is determined in step 702 whether the input character string is a hiragana character string, and if not, step 709.
Go to and use the input character string (Kanji) Kanji buffer 12
After being stored as shown in FIG. 9, the input character string is passed to the predicted character string searching unit 7 in step 710.

On the other hand, if it is determined that the input character string is a Hiragana character string, the input unit 1 proceeds to step 703 to pass the input character string (Hiragana) to the Kanji conversion unit 2. The kanji conversion unit 2 converts the hiragana character string received in step 703 into a kanji character string by referring to the kana character string / kanji dictionary 3, and passes the kanji conversion candidates obtained by the conversion to the used kanji removal unit 11. . If there is a kanji stored in the used kanji buffer 12 among the conversion candidates passed in step 704, the used kanji remover 11 passes the remaining kanji excluding the kanji to the stroke number checker 4.

Here, the concept of removing the same conversion candidate as the Kanji directly input from the input unit 1 by the used Kanji removal unit 11 will be described. When inputting from a handwritten character recognition device, etc., the user can change the character type by inputting the same character string, consciously changing the input character type such as kanji or kana. This will be a hint for the user to predict the character string he wants to input.

For example, when the user inputs all the input character strings in Hiragana, pay attention to the fact that the user consciously inputs in Hiragana instead of inputting in Kanji, and uses this information called Hiragana input. Thus, the character string to be predicted can be limited. Taking the above into consideration, since the character string that the user has input in Kanji directly from the input unit 1 even once has a high probability of being input in Kanji, the character string input in Hiragana is composed of Kanji. When predicting a string,
Even if the predicted character string having the Kanji at the head directly input from the input unit 1 is excluded from the candidates, the number of predicted character string candidates obtained can be reduced without impairing the accuracy of the character string prediction. It is considered that the subsequent processing load can be reduced.

The stroke number investigation unit 4 searches the Kanji / stroke number dictionary 5 for the stroke number of each Kanji conversion candidate received in step 705, and passes it to the character string rearrangement unit 6 together with the Kanji conversion candidate. In step 706, the character string rearrangement unit 6 rearranges the Kanji conversion candidates in descending order of the number of strokes of the Kanji conversion candidates, and then the predicted character string retrieval unit 7 in the rearranged order of the Kanji conversion candidates (corresponding to prioritization). Output to. In step 707, the predictive character string searching unit 7 searches the word dictionary 8 for a predictive character string candidate corresponding to each of the Kanji conversion candidates, and passes it to the character string output unit 9. However, for the kanji character string directly passed from the input unit 1, the corresponding predicted character string candidate is searched from the word dictionary 8 and passed to the character string output unit 9. The predicted character string output unit 9 displays the predicted character string candidates on the display unit 10 in the order of priority in step 708. However, although the predicted character string candidates of the Kanji character string directly passed from the input unit 1 are not prioritized in order of priority, they are also displayed in the same order.

Next, the operation of the above embodiment will be described using a specific example. When the character string “TO” is input as a prediction target, the input unit 1 passes the character string “TO” to the Kanji conversion unit 2. The kanji conversion unit 2 retrieves a conversion candidate for "to" from the kana character string / kanji dictionary 3 as shown in FIG. 3 to obtain "ten, east, throw, party". Used Kanji removal unit 1
For the received conversion candidates (kanji), 1 checks whether or not they are registered in the used kanji buffer 12, and if there is a registered kanji, removes the kanji from the conversion candidates. If there are "ten, party" in the used kanji buffer 12, the used kanji removal unit 11
These are removed and only the conversion candidate “east, throw” is passed to the stroke number investigation unit 4. When the stroke count research unit 4 receives “east, throw”, it searches the kanji / stroke dictionary 5 shown in FIG. 4 for the stroke number of each kanji, and adds the stroke number to the kanji ““ east (8) throw ”. (7) ”is created. The character string rearrangement unit 6 rearranges the conversion candidates according to the number of strokes, here the number in parentheses, and creates "East (8), Throw (7)". The order at this time becomes the output priority of the kanji conversion candidates as it is.

The predictive character string retrieving unit 7 searches the word dictionary 8 shown in FIG. 5 for predictive character string candidates having conversion candidates at the top, in descending order of priority. . When the search is completed for all conversion candidates, they are passed to the character string output unit 9 in the search order. The character string output unit 9 receives “Tokyo, Tokai, pitcher, pitcher” in this order, and displays them on the display unit 10 in the received order.

According to the present embodiment, when predicting a character string composed of Kanji from the input Hiragana character string, the character string which the user tends to input in Kanji directly from the input unit 1 is included at the beginning. Since the character string is not predicted, the number of searched predicted character string candidates can be reduced and the user can easily select and confirm the predicted character string. Moreover, since the probability that the user will input a character string that includes a character string at the beginning, which is apt to be input in Kanji directly from the input unit 1 by the user in Hiragana, is low, the above-described effect can be achieved without impairing the prediction accuracy of the character string. Can be obtained.

FIG. 10 is a block diagram showing still another embodiment of the present invention. The character string prediction device includes an input unit 1, a Kanji conversion unit 2, a kana character string / Kanji dictionary 3, a stroke number investigation unit 4, a Kanji / stroke number dictionary 5, a character string rearrangement unit 6, a predicted character string search unit 7, and a word dictionary. 8, a character string output unit 9, a display unit 10, and a character string selection unit 13.

The functions of the blocks not used in the embodiment of FIG. 1 among the above blocks will be described below. The character string selection unit 13 displays each of the prioritized Kanji conversion candidates on the display unit 10 to allow the user to select a Kanji conversion candidate intended to be input, and the selected Kanji conversion candidate is a predicted character string. It is passed to the search unit 7.

Next, the operation of this embodiment will be described with reference to the flowchart of FIG. Input section 1 is step 1
Enter the character string in 11, and when it becomes a prediction target,
It is determined whether or not the character string input in step 112 is a hiragana character string, and if not, step 11
9, the input character string is passed to the predictive character string searching unit 7,
If it is a hiragana character string, the process proceeds to step 113, and the input character string is passed to the kanji conversion unit 2.

The kanji conversion unit 2 converts the hiragana character string received in step 113 into a kanji character string by referring to the kana character string / kanji dictionary 3, and the kanji conversion candidate obtained by the conversion is the stroke number investigation unit 4. Pass to. The stroke number research unit 4 determines the stroke number of each kanji of the kanji conversion candidates received in step 114.
After searching the stroke number dictionary 5, the stroke number of the Kanji conversion candidate is calculated, and the Kanji conversion candidate and the stroke number thereof are searched for.
Pass to. In step 115, the character string rearrangement unit 6 rearranges the Kanji conversion candidates in descending order of the number of strokes of the Kanji conversion candidates, and then transfers the Kanji conversion candidates to the character string selection unit 13 in the rearranged order (corresponding to prioritization). Output. Character string selection unit 13
In step 116, the conversion candidates converted into kanji are sequentially received from the character string rearranging unit 6, and the received conversion candidates are displayed on the display unit 10 in the order in which they are received. Wait for you to choose. When the conversion candidate displayed on the display unit 10 is the Kanji character string that the user is trying to input, the user presses the select / execute key of the input unit 1 to select the conversion candidate currently displayed. . However, if the input character string is not a kanji character, the user presses the next candidate key of the input unit 1 and the character string selection unit 13 displays the conversion candidate of the next priority.
To display. After that, the character string selection unit 13 passes the conversion candidates selected by the user to the predicted character string search unit 7.
The character string selection unit 13 may collectively display the conversion candidates on the display unit 10, and the user may select a desired conversion candidate from among these.

The predictive character string searching unit 7 searches the word dictionary 8 for a character string having the Kanji conversion candidate passed at step 117 at the beginning, and passes this to the character string output unit 9 as a predictive character string candidate. However, the predicted character string search unit 7 performs step 1
Also for the Kanji character string directly passed from the input unit 1 at 17, a character string having this Kanji character string at the beginning is used in the word dictionary 8
And outputs it to the predicted character string output unit 9. The character string output unit 9 displays the predicted character string candidates passed from the predicted character string search unit 7 in step 118 on the display unit 10 in the order passed. Next, the operation of the above embodiment will be described using a specific example. If "To" is input to the input unit 1 as a prediction target, this character string "To" is converted to the Kanji conversion unit 2
To receive. The kanji conversion unit 2 searches the kana character string / kanji dictionary 3 for a conversion candidate for "to" and obtains "ten, east, throw, party". When the stroke count research unit 4 receives this, "10, east, throw,
The stroke number for each character of the party
Then, the output format in which the number of strokes is added to the kanji conversion candidate, "ten (2), east (8), throw (7), party (10)" is created. The character string rearrangement unit 6 rearranges the received conversion candidates in descending order of the number of strokes to obtain "party (10), east (8), throw (7), ten (2)". The order at this time becomes the output priority of the kanji conversion candidates as it is.

The character string selection unit 13 displays each of "party, east, throw, ten" in the order of reception (priority order), and asks the user to select the kanji intended for input. If “East” is selected here, the character string selection unit 13 passes “East” to the predicted character string search unit 7. The predicted character string search unit 7 searches the word dictionary 8 for a predicted character string candidate having the received Kanji conversion candidate at the head. In this example, the predicted character string search unit 7 searches for “Tokyo, Tokai”, and when the search is completed, passes “Tokyo, Tokai” to the character string output unit 9. The character string output unit 9 displays “Tokyo, Tokai” on the display unit 10 in the order in which they are passed.

According to the present embodiment, the user can select a conversion candidate that the user intended to input from the conversion candidates obtained by converting the hiragana character string that is the prediction target into kanji. Since only the Kanji character string following the selected conversion candidate is predicted, the number of predicted character string candidates to be searched can be narrowed down and the prediction accuracy can be improved. Moreover, since there are few predicted character string candidates to be searched, it is possible for the user to easily select and confirm the predicted character string.

If the usage frequency of the collected character strings is stored in the word dictionary 8, if there are a plurality of predicted character string candidates, the output priority order is rearranged using this usage frequency information. Then, the probability of the predicted character string candidates displayed in the first rank being the character string desired by the user is increased by displaying the predicted character strings on the display unit 10 in order of output priority, and the predicted character string is selected by the user. The confirmation can be simplified and the operability of the device can be improved.

FIG. 12 is a block diagram showing still another embodiment of the present invention. In this example, by adding a used Kanji removal unit 11 and a used Kanji buffer 12 to the configuration of FIG. 11,
It has a configuration in which the configuration shown in the embodiment of FIG. 6 is added in which the user does not predict a character string including a character string at the beginning which tends to be directly input in Chinese characters from the input unit 1. Accordingly, when the character string selection unit 13 causes the user to select a conversion candidate, the user tends to directly input in Kanji from the conversion candidates displayed on the display unit 10 (that is, the character string that predicts this is used. Those who have a low probability of becoming the one desired) can be omitted, and the user can easily select the conversion candidate correspondingly, and the operability of the apparatus can be improved.

The present invention is not limited to the above embodiment. For example, as the kana character string input device, not only a handwritten character recognition device but also a kanji tablet may be used, or kanji conversion. In order to prioritize the candidates, not only the number of strokes but also the frequency of use and the time taken for writing may be used. Further, the predictive character string search unit 7 does not search for all the Kanji conversion candidate character strings, but stops the search when a certain character string is searched for and searches for the predictive character string candidates. You may set the upper limit of number. Furthermore, the character string output unit 9 may not only display the predicted character string candidates, but may pass them to another processing device. Furthermore, if the word written in hiragana character string in the above embodiment is read as a katakana character string and the katakana character string is converted into a kanji character string, the present invention can be easily used for prediction of the katakana character string. It can be applied to obtain the same effect. Further, by installing the character string prediction device of the present invention in a document creation device equipped with a handwriting input device, when the user wants to input a kanji character string with a large number of strokes, only by inputting a part of the reading, Since all Kanji character strings can be input, the efficiency of document creation can be improved.

By the way, in each of the above-mentioned embodiments of the present invention, when the hiragana character string to be predicted is converted into kanji,
When a plurality of conversion candidates are obtained, the function of determining the output priority of these conversion candidates in the order of the number of strokes plays an important role. This is because when a user wants to input a kanji character string, part of it is input in hiragana,
If you want to predict the character string that follows this, it will take more time to enter Kanji in Kanji with more strokes, and there is a tendency to enter in Hiragana. This is based on the idea that the more conversion candidates having a larger number of strokes, the higher the possibility of forming the head portion of the character string that the user wants to input. This idea is also effective when the user inputs characters by handwriting using a doublet and stylus pen, etc., when handwritten characters are input in hiragana and they are recognized and then converted into kana-kanji. The Kanji conversion candidates that the user intended to input are likely to be candidates with a large number of strokes for the above reason. Therefore, the output prioritization of the conversion candidates obtained when the hiragana character string of this example is converted into kanji can be directly applied to the kana-kanji conversion device used in the document creation device equipped with the handwritten character input device. Become.

FIG. 13 is a block diagram showing an embodiment of the kana-kanji conversion device of the present invention made based on the above concept. The kana-kanji conversion device includes an input unit 1, a kanji conversion unit 2, a kana character string / kanji dictionary 3, a stroke number investigation unit 4, a kanji / stroke number dictionary 5, a character string rearrangement unit 6, a conversion candidate output unit 14, and a display unit 10. Composed of. Conversion candidate output unit 1
Reference numeral 4 has a function of displaying kana-kanji conversion candidates on the display unit 10 in the order of output priority, and the functions of other blocks are the same as those of the character string prediction device described above.

Next, the operation of this embodiment will be described with reference to the flowchart of FIG. Input section 1 is step 1
When the character string is input at 41 and temporarily stored in the internal buffer, and when the character string in the buffer becomes a target for kana-kanji conversion by pressing the conversion key or the like, the character string input at step 142 is a hiragana character string. Determine whether or not. As a result, if it is determined that the input character string is not a hiragana character string, the input unit 1 proceeds to step 147 to directly output the input character string in the buffer to the conversion candidate output unit 14. When it is determined that the input character string is a hiragana character string, the input unit 1 proceeds to step 143 and passes the input character string in the buffer to the Kanji conversion unit 2.

The kanji conversion unit 2 converts the hiragana character string received in step 143 into a kanji character string by referring to the kana character string / kanji dictionary 3 (kana kanji conversion), and a kanji conversion candidate obtained by the conversion. Is passed to the stroke count investigation unit 4. The stroke number investigation unit 4 searches the Kanji / stroke number dictionary 5 for the stroke number of each Kanji conversion candidate received in step 144, calculates the stroke number of each Kanji conversion candidate, and adds this stroke number to the Kanji conversion candidate. And passes it to the character string rearrangement unit 6.

The character string rearranging unit 6 rearranges the Kanji conversion candidates received in step 145 in descending order of the number of strokes, and then converts the Kanji conversion candidates in the rearranged order (output priority order). Pass to. Conversion candidate output unit 14
Receives the Kanji conversion candidates in order at step 146,
The information is displayed on the display unit 10 in the order of reception (order of priority). If the Kanji conversion candidate displayed on the display unit 10 is the desired one, the user presses the select / execute key on the input unit 1 to confirm the Kanji conversion candidate. When the candidate key is pressed, the conversion candidate output unit 14 displays the Kanji conversion candidate having the next output priority.

According to the present embodiment, when a hiragana character string is input from a handwritten character input device and the input character string is converted to kana-kanji, the kanji conversion candidates that the user intends to input are output in the higher rank. It is possible to improve the conversion efficiency of Kana and Kanji. By installing the kana-kanji conversion device as shown in FIG. 13 in a document creation device or the like equipped with a handwritten character input device, the character input efficiency can be improved.

[0097]

As described above, claims 1, 2, 3,
According to the ninth, tenth and eleventh inventions, when a kanji character string in which the input hiragana character string is used as the reading of the head part is predicted, the output priority of the predicted character string that the user intends to input , The predicted character string can be easily selected and confirmed by the user.

According to the invention of claim 4 or 12, when it is desired to predict a kanji character string in which the input hiragana character string is the reading of the head part, the number of predicted character strings obtained can be easily narrowed down. It is possible to improve the prediction accuracy of the character string.

According to the invention of claim 5 or 13, even if the obtained kanji conversion candidate is composed of one kanji, the number of strokes as a kanji conversion candidate can be obtained.

According to the invention described in claim 6 or 14, character strings which the user does not intend to input can be extracted from the obtained Kanji conversion candidates, wasteful processing can be omitted, and the obtained predictions can be obtained. By reducing the number of character strings, the user can easily select and confirm the predicted character string.

According to the seventh or fifteenth aspect of the present invention, even if the input character string is not a hiragana character string, a predicted character string for this input character string can be obtained.

According to the eighth or sixteenth aspect of the present invention, it is possible to assign the output priority order of a predicted character string having the same output priority order according to the number of strokes of the Kanji conversion candidate or a predicted character string without an output priority order according to the number of strokes. it can.

According to the invention of claim 19 or 20,
It is possible to improve character input efficiency when a document is created by inputting a kanji character string using a handwritten character input device.

[Brief description of drawings]

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

FIG. 2 is a flowchart showing a character string prediction processing procedure of the apparatus shown in FIG.

FIG. 3 is a diagram showing an example of contents of a kana character string / kanji dictionary used in the embodiment.

FIG. 4 is a diagram showing an example of contents of a kanji / stroke count dictionary used in the embodiment.

FIG. 5 is a diagram showing an example of contents of a word dictionary used in the embodiment.

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

7 is a flowchart showing a character string prediction processing procedure of the apparatus shown in FIG.

FIG. 8 is a diagram showing an output example of a stroke number investigation unit used in the embodiment.

9 is a diagram showing an example of saving the used Kanji character buffer shown in FIG.

FIG. 10 is a block diagram showing still another embodiment of the present invention.

11 is a flowchart showing a character string prediction processing procedure of the apparatus shown in FIG.

FIG. 12 is a block diagram showing still another embodiment of the present invention.

FIG. 13 is a block diagram showing an embodiment of a kana-kanji conversion device of the present invention.

14 is a flowchart showing a kana-kanji conversion processing procedure of the apparatus shown in FIG.

[Explanation of symbols]

1 ... Input part 2 ... Kanji conversion part 3 ... Kana character string / kanji dictionary 4 ... Stroke count part 5 ... Kanji / stroke dictionary 6 ... Character string rearranging part 7 ... Predicted character string search part 8 ... Word dictionary 9 ... Character string Output unit 10 ... Display unit 11 ... Used Kanji removal unit 12 ... Used Kanji buffer 13 ... Character string selection unit 14 ... Conversion candidate output unit

Claims (22)

[Claims] 1. A character string prediction device for searching a character string stored in a dictionary for a character string having an input character string as a head part and outputting the searched character string as a predicted character string. In a character string prediction method for predicting a character string following an input character string, it is determined whether the input character string is a Hiragana character string, and if it is a Hiragana character string, the input character string is converted to Kanji. When the kanji conversion candidates are obtained, the number of strokes of each kanji constituting the kanji conversion candidate is calculated, and then the number of strokes of each kanji constituting one kanji conversion candidate is summed to be the number of strokes of the kanji conversion candidate, and the kanji When there are multiple conversion candidates, the output priority of the predicted character string that has these Kanji conversion candidates at the head is determined by the number of strokes of each Kanji conversion candidate, and then the character string that has these Kanji conversion candidates at the beginning is set as described above. dictionary A character string prediction method, which comprises: searching a character string stored in, and outputting the obtained predicted character string in the output priority order. 2. The character string prediction method according to claim 1, wherein the larger the number of strokes of each Kanji conversion candidate is, the higher the output priority of the predicted character string is. 3. The kanji conversion candidates are rearranged in the output priority order, the predicted character strings of the respective kanji conversion candidates are searched in the rearranged order from the dictionary, and the predicted character strings are output in the searched order. The character string prediction method according to claim 1 or 2. 4. A character string prediction device for searching a character string stored in a dictionary for a character string having an input character string as a head portion and outputting the searched character string as a predicted character string. In a character string prediction method for predicting a character string following an input character string, it is determined whether the input character string is a Hiragana character string, and if it is a Hiragana character string, the input character string is converted to Kanji. When the kanji conversion candidates are obtained, the number of strokes of each kanji constituting the kanji conversion candidate is calculated, and then the number of strokes of each kanji constituting one kanji conversion candidate is summed to be the number of strokes of the kanji conversion candidate, and the kanji When there are multiple conversion candidates, after determining the output priority of these Kanji conversion candidates according to the number of strokes of each Kanji conversion candidate, these Kanji conversion candidates are displayed in this output priority order and wait for the user's selection, That After that, a character string having the Kanji conversion candidate selected by the user at the beginning is searched from the character strings stored in the dictionary, and the obtained predicted character string is output. Prediction method. 5. The character string according to claim 1, wherein when the kanji conversion candidate is composed of one kanji, the number of strokes of this kanji is set as the number of strokes of the kanji conversion candidate. Prediction method. 6. If the input character string is not a hiragana character string, it is further determined whether or not the input character string is a kanji character, and if it is a kanji character, the input character string of this kanji character is stored, and thereafter, From the Kanji candidates obtained by converting the input character string of Hiragana, the same Kanji candidate as the stored Kanji character string is removed, and the processing after determining the output priority order for the remaining Kanji candidates is performed. The character string prediction method according to claim 1, wherein 7. If the input character string is not a Hiragana character string, a predictive character string having this input character string at the head portion is immediately retrieved from the dictionary and output. Or the described string prediction method. 8. When there are a plurality of predicted character strings having the same Kanji conversion candidate or the same input character string at the beginning, the output priority is determined based on a predetermined criterion, and the predicted character strings are placed between the predicted character strings. The character string prediction method according to claim 1, wherein the output is performed only in accordance with the determined output priority order. 9. The character string prediction method according to claim 1, wherein the number of predicted character strings to be searched is set to a preset number or less. 10. A character string predicting device for searching a character string stored in a dictionary for a character string having an input character string as a head part and outputting the searched character string as a predicted character string, Determining means for determining whether or not the input character string is a hiragana character string; and if the determining means determines that the input character string is a hiragana character string, the input character string is converted to kanji and kanji Kana-Kanji conversion means for obtaining conversion candidates, stroke number acquisition means for obtaining the stroke number of each kanji constituting the kanji conversion candidate obtained by this kana-kanji conversion means, and the stroke number of each kanji constituting one kanji conversion candidate The number of strokes calculated by the acquisition means is totaled as the number of strokes of this kanji conversion candidate, and each kanji conversion candidate obtained by the stroke number calculation means when there are a plurality of kanji conversion candidates Comparing means for comparing the number of strokes, determining means for determining the output priority of the predicted character string having the Kanji conversion candidate at the head based on the comparison result of the comparison means, and character having the Kanji conversion candidate at the head portion. Retrieval means for retrieving columns from the character strings contained in the dictionary, display means for displaying the predictive character strings retrieved by the retrieving means in the order of output priority determined by the deciding means, and this display A character string prediction device, comprising: a confirmation unit that allows a user to select an arbitrary character string as a confirmed predicted character string from among the predicted character strings displayed by the means. 11. The character string predicting apparatus according to claim 10, wherein the determining unit determines that the output priority of the predicted character string is higher as the number of strokes of each Kanji conversion candidate is larger. 12. A sorting means is provided for sorting the Kanji conversion candidates in the output priority order determined by the determining means, and the searching means is a predicted character of each kanji conversion candidate in the order sorted by the sorting means. The character string prediction device according to claim 10 or 11, wherein a string is searched from the dictionary, and the output means outputs the predicted character strings in the order searched by the search means. 13. A character string prediction device for searching a character string stored in a dictionary for a character string having an input character string as a head portion and outputting the searched character string as a predicted character string, Judgment means for deciding whether or not the input character string is a hiragana character string, and when the judgment means determines that the input character string is a hiragana character string, the input character string is converted into kanji and kanji Kana-Kanji conversion means for obtaining conversion candidates, stroke number acquisition means for obtaining the number of strokes of each kanji constituting the kanji conversion candidate obtained by this kana-kanji conversion means, and the number of strokes of each kanji constituting one kanji conversion candidate The number of strokes calculated by the acquisition means is totaled as the number of strokes of this kanji conversion candidate, and each kanji conversion candidate obtained by the stroke number calculation means when there are a plurality of kanji conversion candidates Comparing means for comparing the number of strokes, determining means for determining the output priority of the predicted character string having the Kanji conversion candidate at the head based on the comparison result of the comparing means, and the output priority determined by this determining means. Display means for displaying the kanji conversion candidates obtained by the kana-kanji conversion means in order, selection means for allowing the user to select any kanji conversion candidate displayed by this display means, and kanji selected by this selection means Retrieval means for retrieving a character string having a conversion candidate at its head from among the character strings contained in the dictionary, display means for displaying the predicted character string obtained by this retrieval means, and display by this display means A character string prediction device, comprising: a confirmation unit that allows a user to select an arbitrary character string from among the calculated predicted character strings as a confirmed predicted character string. 14. The kanji conversion candidate in the stroke number calculation means is 1
When the character is composed of kanji, the number of strokes of this kanji is set as the number of strokes of the kanji conversion candidate.
The character string prediction device according to any one of 0 to 13. 15. When the determination unit determines that the input character string is not a Hiragana character string, a sub determination unit that determines whether or not the input character string is a Kanji character is provided, and the sub determination unit determines the input character. When it is determined that the column is Kanji, a storage unit that stores the input character string of this Kanji, and a Kanji character string stored in the storage unit from the Kanji candidate obtained by the Kana-Kanji conversion unit. An identifying unit for identifying the same Kanji candidate, a removing unit for removing the Kanji character string identified by the identifying unit from the Kanji candidates, and the determining unit for the remaining Kanji candidates not removed by the removing unit. 15. The character string predicting device according to claim 10, wherein the output priority order is determined for. 16. When the determining unit determines that the input character string is not a Hiragana character string, the searching unit immediately searches the dictionary for a predictive character string having the input character string at the head portion, and the output unit determines 16. The character string prediction device according to claim 10, wherein the searched predicted character string is output. 17. When a plurality of predicted character strings having the same Kanji conversion candidate or the same input character string at the head portion are found as a result of the search by the search means, the output priority order of these predicted character strings is set to a predetermined criterion. 17. The sub-decision means for deciding on the basis of the above is provided, and the output means outputs the output priority order of the predicted character string decided by the sub-decision means in the order of the output priority order. Or the described character string prediction device. 18. The character string prediction device according to claim 10, wherein the search means sets the number of predicted character strings searched from the dictionary within a preset number. 19. A document creating apparatus for predicting a character string to create a document, comprising the character string predicting apparatus according to claim 10, wherein a predictive character string determined by the character string predicting apparatus is provided. A document creating apparatus characterized by inputting in the document. 20. A kana-kanji conversion method in a kana-kanji conversion device for converting an inputted character string into kana-kanji.
After finding the number of strokes of each kanji that composes a kanji candidate obtained by kana-kanji conversion, the number of strokes of each kanji that constitutes one kanji candidate is totaled to be the number of strokes of this kanji candidate, and there are multiple kanji candidates. In this case, the kana-kanji conversion method is characterized in that after the output priority of these kanji conversion candidates is determined by the number of strokes of each kanji candidate, these kanji conversion candidates are displayed in the output priority order. 21. A kana-kanji conversion device for converting an input character string into kana-kanji conversion device, wherein the kana-kanji conversion means for converting the input character string into kanji to obtain kanji conversion candidates, and the kana-kanji conversion device. The number of strokes of each kanji forming the kanji conversion candidate and the number of strokes of each kanji forming one kanji conversion candidate obtained by the stroke number acquiring means Stroke number calculating means, a comparing means for comparing the number of strokes of each kanji conversion candidate obtained by the stroke number calculating means when there are a plurality of kanji conversion candidates, and the kanji conversion candidate based on the comparison result of the comparing means. And a kanji conversion candidate obtained by the kana-kanji conversion means in the order of the output priority order determined by this determination means. A kana-kanji conversion device comprising: display means for displaying. 22. A kana according to claim 21, further comprising: a handwriting input means for inputting characters by handwriting, wherein the character string input by the handwriting input means is converted into kana by kanji conversion and input. Equipped with a kanji conversion device,
When the input character string is hiragana, the kana-kanji conversion device is used when converting the hiragana character string into kana-kanji characters.