JP2010061407A - Character input device, method and program, and portable telephone device - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To present effective prediction candidates for accessory words in the past form in a text in the past tense. <P>SOLUTION: A character input device includes: a prediction word database 101 for storing the candidates of words for prediction via a cursive form of Japanese syllabary; a connection prediction accessory word database 102 which stores the candidates of accessory words for connection prediction and to which information of past tense decision is applied; a tense term database 103 for storing words showing the past; a tense priority mode storage memory 104 for storing setting information of the past priority mode; a candidate output storage memory 105 for storing the candidates of the accessory words for connection prediction pertinent to the words for prediction; and a prediction input device 109 for making a display device display the candidates of the words for prediction on the basis of the retrieval of the words for prediction in the middle of input, and for making the storage memory for candidate output store the result on the basis of the retrieval of the accessory words for connection prediction, and for rearranging and storing the candidates decided to be the past tense among the candidates of the stored accessory words for prediction in the high order sequence when the setting information of the past priority mode is stored, and for making the display device display the stored candidates. <P>COPYRIGHT: (C)2010,JPO&INPIT

Description

  The present invention relates to a character input device that predicts and inputs accessory word candidates in accordance with part-of-speech information given to each input word. In particular, the present invention relates to a character input device, a method, a program, and a mobile phone device that are capable of preferential input by predicting connected word candidates in accordance with the currently input tense.

Generally, in a character input method in a mobile phone device, a line of 50 sounds is assigned to each of the numeric keys “1” to “0”, and the character on the assigned line changes depending on the number of times the key is pressed. Adopted.
In this method, in order to press “O”, it is necessary to press “1” five times. Therefore, there is a problem that the number of key presses increases explosively as the number of input characters increases.

Therefore, in order to reduce the number of times of pressing, a means of predicting a so-called prefetched word in which the hiragana being input matches the front of the hiragana of the word to be input is becoming common.
Further, the prediction input often has a function of predicting a word following the word from the input words (hereinafter referred to as connection prediction).

In this connection prediction, connection adjunct word candidates are determined according to the part of speech information given to each word.
Therefore, fixed adjunct words are linked and displayed as prediction candidates by the part of speech information of the selected and confirmed words as shown in the following example.
FIG. 9 is a diagram for explaining a specific example of character input as a premise of the present invention. As shown in this figure (a), when hiragana “ta” is entered and displayed on the text creation screen, a plurality of prefetched words starting with “ta” are fixed in the order of frequency of use in the lower column of the text creation screen. Displayed.

As shown in FIG. 5A, “fun” is selected from the lower column of the text creation screen, and the word “fun” is input to the text creation screen. As shown in this figure (b), when the word “fun” is entered on the text creation screen, the word “fun” entered in the lower column of the text creation screen is fixed in order of frequency of use. Is displayed.
However, no matter what kind of sentences are created, since the adjunct candidates are displayed in a fixed order, even when creating sentences about past cases, the present tense There is a problem in that it is troublesome to select a past tense adjunct in past tense sentences because it is displayed at the top and past tense adjuncts are displayed below and may not be displayed unless scrolled.

Therefore, in past tense sentences, it becomes a problem to dynamically connect to past tense adjunct words and rearrange them so that the rank of the adjunct word candidates becomes higher and present more effective prediction candidates.
Conventional techniques related to character input, which is a premise of the present invention, include the following.
Conventionally, learning the past misrecognition results, and reflecting the learning results of character candidate groups and predicted character string groups, which are recognition results in subsequent character input, makes it easy for the user to intend characters and character strings. In a method in which a candidate group of recognized characters is displayed in a sequence according to priority order, and a user selects an arbitrary candidate character from among them and inputs a character, When a character other than the first candidate is selected, the selected character candidate (correct character), the first candidate character (misrecognized character), and a candidate character group are registered in the recognition learning dictionary. When displaying the recognition result, there is one that determines the arrangement of each candidate character group after referring to the recognition learning dictionary and taking into account the past erroneous recognition results (for example, see Patent Document 1).

However, in the above-mentioned Patent Document 1, those that are dynamically connected to past tense in the past tense and rearranged so that the rank of the predecessor candidate is higher and present more effective prediction candidates. Not listed.
Also, in the past, it was not known what word to select when performing a refined search in the search, and the search engine was repeated repeatedly, increasing the burden on the search engine. The information retrieval device refers to the keyword management file to check whether the keyword is managed. If there is a search history, the information retrieval device uses the corresponding keyword and the keyword from the search history database. As a result, the user knows the keywords that can be easily searched for one keyword continuously without actually performing the search. In addition, it is possible to know the predicted value of the search result using the keyword, and to obtain the desired search result. Can reach quickly, yet also being capable of reducing the burden on the search engine is (e.g., see Patent Document 2).

However, in the above-described Patent Document 2, as described above, the past tense in the past tense is dynamically connected to the past tense so that the rank of the auxiliary word candidates is rearranged, and more effective prediction candidates. The thing which presents is not described.
Further, conventionally, in order to support an input operation in an information processing terminal smoothly, a server apparatus that holds input dictionary information that associates a character string with a character string candidate that can be input following the character string, A plurality of information processing terminals are connected, and each information processing terminal acquires and stores input dictionary information from the server device, retrieves and presents character string candidates for the input character string from the input dictionary information, Character string candidates corresponding to the character string candidates selected from the presented character string candidates are retrieved from the input dictionary information and presented, the selected contents of the character string candidates are notified to the server apparatus, and the server apparatus displays the notified selection contents. The input dictionary information is updated by accumulating the selection frequency of the character string candidate in the input dictionary information based on the result and setting the priority order of the character string candidate based on the accumulated result. That (for example, see Patent Document 3).

However, in the above-mentioned Patent Document 3, as described above, the past tense in the past tense is dynamically connected to the past tense so that the rank of the subordinate candidate is rearranged, and a more effective prediction candidate. The thing which presents is not described.
In addition, in the conventional technology for assisting input by predicting the character string that follows in the course of inputting text, a method for inputting an ellipsis and performing forward matching has a heavy burden at the time of character input and is selected in the past. In the technique for learning the candidates, various problems have occurred that the first input word cannot be handled. Therefore, in the present invention, in predicting characters, field information is assigned for each kanji and the following is followed. Among character string candidates, by narrowing down candidates based on the field information assigned to each kanji and displaying candidate character strings, there are also those that perform character input by prediction with a high accuracy rate of prediction candidates (for example, (See Patent Document 4).

However, in the above-mentioned Patent Document 4, as described above, the past tense in the past tense is dynamically connected to the past tense so that the rank of the auxiliary word candidates is rearranged, and a more effective prediction candidate. The thing which presents is not described.
Conventionally, in order to complement even a plurality of words or characters omitted in a sentence, an input device for inputting characters, a display device for displaying the input characters, and conversion and editing of the input characters are performed. A document processing device having a processing device for performing processing, the processing device including a word dictionary storing words and their usage frequency, a transition dictionary between words describing transition information between words and transition probabilities between words, Candidate word extraction means for extracting candidate words omitted by referring to the word dictionary from characters before and after the ellipsis included in the input sentence, and among candidate words extracted based on the transition dictionary between words Some of them are configured by a determination means for determining one word from (see, for example, Patent Document 5).

  However, in the above-mentioned patent document 5, as described above, the past tense in the past tense is dynamically connected to the past tense so that the rank of the auxiliary word candidates is rearranged, and more effective prediction candidates. The thing which presents is not described.

JP 2000-348141 A JP 2003-108594 A JP 2006-155322 A Japanese Patent Laid-Open No. 10-269210 Japanese Patent Application Laid-Open No. 11-282832

  Therefore, in view of the above-mentioned problems, the present invention dynamically rearranges the past word ancillary words in the past tense so that the ranking of the adjunct word candidates is higher, and makes a more effective prediction candidate. An object is to provide a character input device, a method, a program, and a mobile phone device to be presented.

  In order to solve the above problems, the present invention provides a character input device that predicts a candidate for a predictive auxiliary word connected to each word, and displays a created sentence and input hiragana on a sentence creation screen. A display unit for displaying a prediction word candidate for which a prefetching input is predicted in the column or a prediction auxiliary word candidate connected to the selected prediction word; a prediction word candidate displayed on the display device; and connection prediction Predictive words selected based on candidate adjuncts, an e-mail creation device that creates e-mail text from connection predictive adjuncts, and multiple predictions in the order of frequency of use via hiragana in the middle of input Prediction word database for storing candidate words and multiple connection prediction candidate words in order of frequency of use with the prediction word as a medium. For connection prediction appendix database to which information for judging whether or not, tense term database that stores multiple past words as past priority candidates, and sentences that contain past words Temporary priority mode storage memory for saving past priority mode setting information and a plurality of connection prediction ancillary word candidates passed from the connection prediction ancillary word database, and a plurality of connection prediction ancillary word candidates On the other hand, a candidate output storage memory for rearranging and storing candidate words for connection prediction determined to be past tense in a higher rank and the word in the sentence created by the e-mail creating device as a medium If there is a corresponding past priority candidate based on the search of the tense term database, the past priority mode setting information is stored in the tense priority mode storage memory and entered. Based on the search of the prediction word database through intermediate hiragana, the corresponding prediction word candidates are displayed on the display device, and the connection using the prediction word selected by the e-mail creation device as a medium Predictive adjunct word candidates linked to the search of the predictive adjunct database are stored in the candidate output storage memory, and the past priority mode setting information is stored in the tense priority mode storage memory Prediction adjunct candidates determined to be past forms out of the prediction adjunct candidates stored in the candidate output storage memory are rearranged in higher order and stored, and stored in the candidate output storage memory There is provided a character input device comprising: a prediction input device for displaying a candidate for a prediction accessory word on the display device.

Further, when a punctuation mark including “.”, “,”, “?”, “!” Is input, the prediction input device stores the setting information of the past priority mode in the tense priority mode storage memory. Release it.
Further, the prediction word database further stores a plurality of prediction word candidates having a past tense, and is the past tense for each of the plurality of prediction word candidates having a past tense? And the prediction input device stores the prediction word stored in the candidate output storage memory when the setting information of the past priority mode is stored in the tense priority mode storage memory. The candidates for the prediction word determined to be the past tense among the candidates are rearranged in the highest order and stored, and the prediction word candidates stored in the candidate output storage memory are displayed on the display device.

  Further, the tense term database stores a plurality of words representing the future as future priority candidates, and the tense priority mode storage memory stores setting information of the future priority mode for a sentence that includes words representing the future and is being created. The predictive input device uses the tense priority mode storage memory when a future priority candidate exists based on a search of the tense term database using a word in a sentence created by the e-mail creation device as a medium. The setting information of the future priority mode is stored on the display device, and the prediction word candidate corresponding to the prediction word database search based on the hiragana in the middle of the input is displayed on the display device. Candidates for prediction adjuncts connected based on a search of the connection prediction adjunct database using the selected prediction word as mediator When the setting information of the future priority mode is stored in the tense priority mode storage memory, it is determined as the future form among the prediction adjunct candidates stored in the candidate output storage memory. The prediction adjunct candidates are rearranged in a higher order and stored, and the prediction adjunct candidates stored in the candidate output storage memory are displayed on the display device.

  Furthermore, the present invention relates to a word input method for predicting a candidate for a predictive auxiliary word connected to each word, and a word in a sentence created for a tense term database that stores a plurality of words representing the past as past priority candidates. Searching as an intermediary, and if there is a corresponding past priority candidate, storing the past priority mode setting information in the tense priority mode storage memory, and multiple predictions in descending order of usage frequency through hiragana in the middle of input Searching for a prediction word database that stores candidate words for use through hiragana in the middle of input and displaying the corresponding prediction word candidates in the lower column of the sentence creation screen of the display device; Information for storing multiple connection prediction ancillary candidates as an intermediary in order of frequency of use and determining whether each connection prediction appendix is a past tense Searching the given connection prediction adjunct database using the prediction word selected by the display device as a medium, and storing the corresponding prediction adjunct word candidates in a candidate output storage memory; When the setting information of the past priority mode is stored in the priority mode storage memory, the prediction auxiliary word candidates determined to be the past form among the prediction auxiliary word candidates stored in the candidate output storage memory are selected. Rearranging and saving in a higher order, and displaying prediction adjunct word candidates saved in the candidate output storage memory in the lower column of the sentence creation screen of the display device; and a sentence creation screen of the display device Creating an e-mail sentence on the sentence creation screen of the display device by selecting a prediction word candidate or a connection prediction auxiliary word candidate displayed in the lower column of It provides a character input method comprising and.

  Further, the present invention provides a character input program that predicts candidate words for prediction and connected to each word, and stores words in a sentence created for a tense term database that stores a plurality of words representing the past as past priority candidates. Searching as an intermediary, and if there is a corresponding past priority candidate, a procedure for saving the setting information of the past priority mode in the tense priority mode storage memory and a plurality of predictions in the order of frequency of use through the hiragana in the middle of input A procedure for searching the prediction word database for storing candidate words using hiragana in the middle of input and displaying the corresponding prediction word candidates in the lower column of the sentence creation screen of the display device, and the prediction word To store multiple connection prediction ancillary candidates as an intermediary in order of frequency of use, and to determine whether each connection prediction appendix is a past tense A procedure for searching for a word for prediction selected by the display device with respect to a connection prediction auxiliary word database to which information is given and storing the corresponding prediction auxiliary word candidate in a candidate output storage memory; When the setting information of the past priority mode is stored in the tense priority mode storage memory, the prediction auxiliary word determined to be the past form among the prediction auxiliary word candidates stored in the candidate output storage memory The candidate is rearranged in a higher rank order and saved, and the prediction adjunct word candidate saved in the candidate output storage memory is displayed in the lower column of the sentence creation screen of the display device, and the sentence of the display device A procedure for creating an e-mail sentence on the sentence creation screen of the display device by selecting a prediction word candidate or a connection prediction auxiliary word candidate displayed in the lower column of the creation screen. Providing a character input program, characterized in that to obtain.

  Furthermore, the present invention provides a mobile phone device that creates an e-mail by predicting candidate words for prediction linked to each word, and displays the created text and input hiragana on the text creation screen. A display unit for displaying a prediction word candidate for which a prefetching input is predicted in the column or a prediction auxiliary word candidate connected to the selected prediction word; a prediction word candidate displayed on the display device; and connection prediction Predictive words selected based on candidate adjuncts, an e-mail creation device that creates e-mail text from connection predictive adjuncts, and multiple predictions in the order of frequency of use via hiragana in the middle of input Prediction word database for storing candidate words and multiple connection prediction candidate words in order of frequency of use through the prediction word as an intermediate, and each connection prediction accessory word in past form For connection prediction appendix database to which information for determining whether or not is present, tense term database that stores a plurality of words representing the past as past priority candidates, and a sentence that includes a word representing the past and is being created A temporal priority mode storage memory for storing setting information of the past priority mode, and a plurality of connection prediction ancillary words passed from the connection prediction ancillary word database, and a plurality of connection prediction ancillary word candidates A candidate output storage memory that rearranges the candidate words for connection prediction determined to be past tense in a higher order and stores them, and a word in the sentence created by the e-mail creation device If there is a corresponding past priority candidate based on the search of the tense term database, the past priority mode setting information is stored in the tense priority mode storage memory, The prediction word candidate corresponding to the search word database search based on hiragana in the middle of power is displayed on the display device, and the prediction word selected by the e-mail creation device is used as a medium. When a candidate for a predictive adjunct is stored in the candidate output storage memory based on the search of the connection predictive adjunct database, and setting information of the past priority mode is stored in the tense priority mode storage memory The prediction adjunct word candidates determined to be past forms among the prediction adjunct candidates stored in the candidate output storage memory are rearranged in a higher order and stored, and stored in the candidate output storage memory And a prediction input device that displays the candidate for the prediction auxiliary word on the display device.

  As described above, according to the present invention, search is performed using words in a sentence created for a tense term database that stores a plurality of words representing the past as past priority candidates, and there is a corresponding past priority candidate. For the prediction word database that stores the setting information of the past priority mode in the tense priority mode storage memory, and stores a plurality of prediction word candidates in the order of frequency of use through the hiragana in the middle of input Searches using hiragana in the middle of input and displays corresponding prediction word candidates in the lower column of the text creation screen of the display device, and uses a plurality of connected prediction candidate words frequently using the prediction word as a medium It is saved in order, and the connection prediction ancillary database is given on the display device to which information for judging whether it is a past tense for each connection prediction ancillary word. If the candidate for the associated predictive adjunct is searched for in the candidate output storage memory and the setting information of the past priority mode is stored in the tense priority mode storage memory, the candidate is searched. Prediction adjunct word candidates determined to be the past tense among prediction adjunct candidates stored in the output storage memory are rearranged into higher ranks and stored, and the prediction adjunct words stored in the candidate output storage memory Display adjunct word candidates in the lower column of the text creation screen of the display device, and select the prediction word candidate or candidate for connection prediction adjunct word displayed in the lower column of the text generation screen of the display device. Since the e-mail text is created on the text creation screen of the device, a fixed attached word is connected and displayed as a prediction candidate by the part of speech information of the selected and confirmed word in the conventional prediction input method. According to the present invention, by giving information as to past tense to the candidate for connection prediction appendix, the past appendix candidate for the past tense that fits the past tense of the input sentence is given priority. Can be displayed. Also, when a punctuation mark is entered, the past priority mode setting is automatically canceled because there is a high possibility that the past tense will not continue to be displayed preferentially as a sentence flow. Thus, a more comfortable prediction input function can be provided. Furthermore, the present invention can be applied not only to a prediction ancillary word but also to a prediction word. Furthermore, the present invention can be applied not only to a past tense but also to a future tense.

Hereinafter, embodiments of the present invention will be described with reference to the drawings.
FIG. 1 is a block diagram showing a schematic configuration of a character input device according to the present invention. As shown in the figure, a character input device 100 provided in a mobile phone device includes a prediction word database 101, a connection prediction auxiliary word database 102, a tense term database 103, a tense priority mode storage memory 104, and a candidate output storage memory. 105, a display device 107, an input means 106, a database search device 108, a prediction input device 109, and an e-mail creation device 110, which are dynamically connected to past tense adjuncts in sentences and rank higher in the ranking of adjunct word candidates. The program is executed so that more effective prediction candidates are presented.

The database search device 108, the prediction input device 109, and the e-mail creation device 110 are constituted by a CPU (central processing unit).
The prediction word database 101 is composed of a ROM (read-only memory), and words that are predicted to be prefetched through hiragana in the middle of input are fixed as a plurality of prediction word candidates in descending order of frequency of use. save.

The connection prediction adjunct database 102 is composed of ROM, and the adjuncts such as particles and auxiliary verbs that are connected to the word through the word are fixed and stored as a plurality of connection prediction adjunct candidates in descending order of frequency of use. Each connection prediction appendix is given information for determining whether it is a past tense.
The tense term database 103 is composed of a ROM, and stores a plurality of words representing the past as past priority candidates.

The tense priority mode storage memory 104 is composed of a RAM (Random Access Memory) and includes a word representing the past during sentence creation, and the past priority mode is set for the sentence being created. Save the priority mode setting information.
The candidate output storage memory 105 is composed of a RAM, stores a plurality of connection prediction ancillary word candidates passed from the connection prediction ancillary word database 102, and stores the past tense for a plurality of connection prediction ancillary word candidates. It has a function of rearranging the candidates for connection prediction adjunct words determined as follows in the higher order and storing them.

The input means 106 is composed of a numeric keypad, inputs an email reading created by the email creation device 110, and selects a predicted word candidate and a predicted associated word candidate.
The display device 107 includes an LCD (liquid crystal display unit), and displays input hiragana, predicted word candidates, predicted attached word candidates, and sentences created by the e-mail creation device 101.

  The database search device 108 searches the prediction word database 101 for word candidates whose prefetch input is predicted through the hiragana in the middle of input, and predicts the word input to the connection prediction word database 102. A search is made for a candidate for an attached word, and a search is made as to whether a word representing the past is stored in the words in the sentence created for the tense term database 103.

  The prediction input device 109 displays the predicted word candidates on the display device 107 by the search of the database search device 108, and further stores the prediction auxiliary word candidates connected to the input words in the candidate output storage memory 105. In addition, when a word representing the past is stored as a word in the sentence created in the tense term database 103 and displayed on the display unit 107, the setting information of the past priority mode is stored in the tense priority mode storage memory 104. If the setting information of the past priority mode is stored in the tense priority mode storage memory 104, the prediction attachment determined as the past tense among the prediction auxiliary word candidates stored in the candidate output storage memory 105 The word candidates are rearranged in the higher order and stored, and displayed on the display device 107.

  Note that the prediction input device 109 stores the tense priority mode when punctuation marks such as “.”, “,”, “?”, “!”, Etc. are input during the creation of the email text by the email creation device 110. The storage of the setting information of the past priority mode in the memory 104 is canceled. As a result, when the sentence is completed, the past tense priority process is not continued in the subsequent prediction input.

The e-mail creating apparatus 110 selects the prediction word and the connection prediction attachment selected when the input unit 106 selects the prediction word candidate and the connection prediction attachment word candidate displayed on the display device 107. Create e-mail text in words.
FIG. 2 is a diagram for explaining a storage example of the connection prediction appendix database 102 and a storage example of the candidate output storage memory 105 in FIG. The example of the connection prediction appendix database 102 shown in FIG. 9A is the same as the example displayed in FIG. 9B, and a plurality of connection prediction appendices “I”, “K”,. “Too many” are usually arranged and fixed in order of use frequency and stored, and the past tense column is provided in the past tense column in the past tense column in the past tense column in the appendix database 102 for connection prediction. For example, in the case of an adjunct, the flag “1” is past for “Did”, “Did it?”, “Did not?”, “Did not?” It is given and saved as shape information.

  In the candidate output storage memory 105 shown in FIG. 5B, the connection prediction stored in the connection prediction adjunct database 102 when the information of the past priority mode setting is not stored in the tense priority mode storage memory 104. The adjuncts “I”, “ku”,..., “Too much” are stored as they are in the order of frequency of use. Prediction adjunct database 102 past tense column with “1” flag connected prediction adjunct, for example, “Did”, “Yes”, “Did not”, “Did not Rearrange the ranks of “Ta” and “Too much” to the top and save.

  FIG. 3 is a diagram for explaining an example of saving the tense term database 103 in FIG. As shown in the figure, the tense term database 103 includes, for example, “Kino”, “Ototoi”, “Sakio Totoi”, “Senshu”, “Sengetsu”, “Kyonen”, etc. as words representing the past. The display “Yesterday”, “Yesterday”, “Yesterday yesterday”, “Last week”, “Last month”, “Last year”, etc. are also saved for the word read.

FIG. 4 is a diagram for explaining an example of creating past tense sentences. As shown in FIG. 5A, when the sentence “Last Week Movie” is created on the sentence creation screen of the display device 107, the word “Last Week” in the created sentence is a word representing the past. The mode storage memory 104 stores setting information for the past priority mode.
If there is an input of hiragana “ta” after “last week's movie” on the text creation screen, candidates for prefetched prediction words are displayed in the lower column of the text creation screen of the display device 107 as in FIG. When “fun” is selected from the candidates of the pre-read prediction words in the lower column of the text creation screen of the display device 107, “fun” is created after the text of “last week movie” as shown in FIG. Entered on the screen.

  When the prediction word “fun” is input, the prediction input device 109 stores the setting information of the past priority mode in the tense priority mode storage memory 104, and therefore the prediction auxiliary word stored in the candidate output storage memory 105. Among the candidates, prediction adjunct word candidates determined to be past tense are rearranged in a higher order and stored, and displayed on the display device 107. This makes it easy to input "I enjoyed", "Looked fun", "Not fun", "Not fun", "I was too much fun", and so on.

FIG. 5 is a flowchart for explaining a series of operation examples of the character input device shown in FIG. As shown in the figure, in step S201, “reading” of hiragana is input from the input means 106 in the creation of an email by the email creation device 110.
In step S <b> 202, when input of “reading” of hiragana is performed, the prediction input device 109 causes the database search device 108 to search for a prediction word candidate as a prefetched word candidate.

In step S <b> 203, prediction word candidates are displayed on the display device 107 as prediction candidates searched from the prediction word database 101 via the database search device 108.
In step S204, it is determined whether one prediction word is selected by the input unit 106 from the prediction word candidates displayed on the display device 107 and the input word is determined. If there is no word to be determined, the process is terminated, and further, the continuation of the “reading” in step S201 is input by the input unit 106 and the above process is continued or the input unit 106 is not used. The desired word is directly input from.

When the input word is determined in step S205, the predictive input device 109 searches the tense term database 103 with the database search device 108 for words in the e-mail text created by the e-mail creation device 110. Let
In step S206, the prediction input device 109 compares the words in the e-mail text created by the e-mail creation device 110 with the past priority candidates stored in the tense term database 103, and there is a matching past priority candidate. It is determined whether or not to do. If not, the process proceeds to step S208.

If there is a matching past priority candidate in step S207, the prediction input device 109 stores the setting information of the past priority mode in the tense priority mode storage memory 104.
In step S <b> 208, the prediction input device 109 connects to the database search device 108 for the word determined to be input, and searches the prediction auxiliary word database 102.
In step S <b> 209, the prediction input device 109 stores, in the candidate output storage memory 105, the connection prediction auxiliary word candidates searched in the connection prediction auxiliary word database 102 for the word whose input is determined.

In step S <b> 210, the prediction input device 109 determines whether the past priority mode setting information is stored in the tense priority mode storage memory 104. If not, the process proceeds to step S212.
In step S211, if the past priority mode setting information is stored, the prediction input device 109 determines that the connection prediction word candidate stored in the candidate output storage memory 105 is the past tense. Predictive adjunct word candidates are sorted and saved in higher order.

In step S <b> 212, the prediction input device 109 causes the display device 107 to display the connected prediction auxiliary word candidates stored in the candidate output storage memory 105 in the stored order.
In step S <b> 213, it is determined whether one of the candidate words for connection prediction displayed on the display device 107 is selected as an e-mail text created by the e-mail creation device 110. If no adjunct word candidate is selected, the process ends, and the desired adjunct word is directly input by the input means 106.

In step S214, a candidate for an adjunct word is selected, and the next input to the e-mail creation text of the e-mail creation device 110 is past punctuation marks such as “.”, “,”, “?”, “!”, Etc. It is determined whether the input is for canceling the priority mode setting. If it is not an input to cancel, the process proceeds to step S216.
In step S215, when the input is to cancel the setting of the past priority mode, the prediction input device 109 cancels the storage of the setting information of the past priority mode stored in the tense priority mode storage memory 104.

In step S216, it is determined whether or not the sentence has been created. If it is completed, the process ends. If not completed, the process returns to step S201.
Therefore, in the conventional predictive input method, a fixed ancillary word is displayed as a connected prediction candidate based on the part of speech information of a selected and confirmed word, but according to the present invention, a connected predicted ancillary word candidate is a past tense. By giving such information, it becomes possible to preferentially display past tense associated word candidates that match the past tense of the input sentence. Also, when a punctuation mark is entered, the past priority mode setting is automatically canceled because there is a high possibility that the past tense will not continue to be displayed preferentially as a sentence flow. Thus, a more comfortable prediction input function can be provided.

As described above, the rearrangement of the order of the candidates for the connection prediction appendix has been described. However, the past tense at the time of the prefetch prediction can be used as follows.
FIG. 6 is a diagram for explaining a storage example in consideration of the past tense of the prediction word database 101 of FIG. The example of the prediction word database 101 shown in this figure is the same as the example displayed in FIG. 9B, and a plurality of prediction word candidates “just”, “very”, “eat”,. , “Taxi”, “It was hard”, “I ate”, “It was fun”,…, “Occasionally” were usually arranged and stored in order of frequency of use, and each prediction was stored in the prediction word database 101. If you have a past tense column for candidate words, and the past tense column is a past tense prediction word, for example, “It was hard”, “I ate”, “I had fun” ..., a flag “1” is assigned to “sometimes” as past tense information and stored.

  FIG. 7 is a diagram for explaining an example of creating past tense sentences using the prediction word database 101 of FIG. As shown in this figure, when the sentence “Last Week Movie” is created on the sentence creation screen of the display device 107, the word “Last Week” in the created sentence is a word representing the past. In 104, setting information of the past priority mode is stored. If there is an input of hiragana “ta” after “last week's movie” on the text creation screen, candidates for prefetched prediction words in the lower column of the text creation screen of the display device 107 are “hard”, “eat”, Candidates for prediction words displayed in the prediction word database 101, such as “It was fun”, “Sometimes”, “Just”, “Hard”, “Eat”,…, “Maybe” The candidates of predicted words determined to be past tense are rearranged in the higher rank and displayed. This makes it easy to input “I enjoyed it”.

  To simplify the explanation, I limited it to "It was hard", "I ate", "It was fun", ... "Sometimes", but "It was very hard" for "It was hard" "It was not so hard", "It was too hard", and "I ate it", "I didn't want to eat", "I ate it" "I didn't like it" "," I ate too much ", and" It was fun "," It wasn't fun "," It didn't seem to be fun "," It was too much fun " ,… You may add “it sometimes seems to be”, “sometimes gone”, “it seems to be gone”, or “sometimes” to “sometimes”.

  In this case, if only the input of hiragana “ta” is displayed, the above prediction word candidates are displayed at one time. For example, if input is made up to hiragana “tano”, “Looked fun”, “Not fun”, “Not fun”, “Too fun” are displayed at the top. Others are the same. In this way, the versatility of the present invention can be expanded.

In the above description, the candidate for the predicted ancillary words related to the past tense is dynamically rearranged in priority, but it can be explained as follows for the future tense as well. is there.
FIG. 8 is a diagram for explaining an example of creating a futuristic sentence. As shown in the figure (a), when the text “next week's movie” is created on the text creation screen of the display device 107, the word “next week” in the created text is a word representing the future, so the tense priority is given. The mode storage memory 104 stores setting information for the future priority mode. The tense term database 103 for storing words representing the future includes “morrow”, “morrow” (morning), “morrow” (morning), “next week (raishu)”, “next year (rai) Nene) "etc. are saved.

  If there is an input of “tana” after “next week's movie” on the text creation screen, prefetched prediction word candidates are displayed in the lower column of the text creation screen of the display device 107 in the same order as in FIG. . When “fun” is selected from the prediction word candidates in the lower column of the text creation screen of the display device 107, “fun” is created after the text of “next week's movie” as shown in FIG. Entered on the screen.

  When the prediction word “fun” is input, since the setting information of the future priority mode is stored in the tense priority mode storage memory 104 in the prediction input device 109, the prediction auxiliary word stored in the candidate output storage memory 105. Among the candidates, “Mimi”, “I am looking for”, “I am”, “I will”, “Mu should be”, etc. are ranked higher as candidates for the predictive adjunct The data is rearranged and saved, and displayed on the display device 107. This makes it easy to input “I'm looking forward”, “I ’m looking forward”, “I ’m looking forward”, “I ’m going to have fun”, “I should have fun”, etc. In this way, the versatility of the present invention can be extended to future tenses.

  In the above description, the character input device is provided in the mobile phone device. However, the present invention is not limited to this and can be used in a PDA (personal digital assistant) and a PHS (simple mobile phone).

It is a block diagram which shows schematic structure of the character input device which concerns on this invention. FIG. 2 is a diagram for explaining a storage example of a connection prediction appendix database 102 and a storage example of a candidate output storage memory 105 in FIG. 1. It is a figure explaining the preservation | save example of the tense term database 103 in FIG. It is a figure explaining the creation example of the past tense sentence. It is a flowchart explaining a series of operation examples of the character input device in FIG. It is a figure explaining the example of a preservation | save which considered the past tense of the prediction word database 101 of FIG. It is a figure explaining the example of creation of the past tense sentence using the word database for prediction 101 of FIG. It is a figure explaining the example of preparation of the future form sentence. It is a figure explaining the specific example of the character input used as the premise of this invention.

Explanation of symbols

DESCRIPTION OF SYMBOLS 100 ... Character input device 101 ... Prediction word database 102 ... Connection prediction appendix database 103 ... Temporal term database 104 ... Temporal priority mode storage memory 105 ... Candidate output storage memory 106 ... Input means 107 ... Display device 108 ... Database search Device 109 ... Predictive input device 110 ... E-mail creation device

Claims (7)

In a character input device that predicts a candidate for an adjunct word for prediction connected for each word,
The created text and the input hiragana are displayed on the text creation screen, and the prediction word candidates predicted to be prefetched or predicted adjunct words connected to the selected prediction word are displayed in the lower part of the text creation screen. A display unit to
A prediction word displayed on the display device, a prediction word selected on the basis of a connection prediction ancillary word candidate, an e-mail creation device that creates an e-mail sentence from the connection prediction ancillary word; and
Prediction word database that stores a plurality of prediction word candidates in order of frequency of use through hiragana in the middle of input,
Prediction words are stored in order of frequency of use, and predictions are given to each connection prediction auxiliary word with information for determining whether it is a past tense. An adjunct database for
A tense term database that stores multiple past words as past priority candidates,
A tense priority mode storage memory that stores past priority mode setting information for a sentence that includes a word representing the past and is being created;
A plurality of connection prediction ancillary word candidates passed from the connection prediction ancillary database are stored, and connection prediction word adjunct candidates that are determined to be a past tense with respect to a plurality of connection prediction ancillary word candidates are stored. Candidate output storage memory that is sorted and saved in the higher rank,
Setting information of the past priority mode in the tense priority mode storage memory when there is a corresponding past priority candidate based on the search of the tense term database using the words in the sentence created by the e-mail creation device The prediction word candidate corresponding to the prediction word database search based on the hiragana in the middle of input is displayed on the display device, and the prediction word selected by the e-mail creation device is displayed. Predictive adjunct word candidates connected to the connection prediction adjunct database search as a mediator are stored in the candidate output storage memory, and past priority mode setting information is stored in the tense priority mode storage memory. If there are prediction adjunct candidates stored in the candidate output storage memory, the prediction adjunct candidates determined to be the past tense are ranked higher Rearranged and stored, the character input device characterized in that it comprises a predictive input device for displaying the candidates of the candidate prediction appendage words stored in the output storage memory to the display device. When a punctuation mark including “.”, “,”, “?”, “!” Is input, the prediction input device causes the tense priority mode storage memory to cancel the saving of the setting information of the past priority mode. The character input device according to claim 1, wherein: The prediction word database further stores a plurality of prediction word candidates having a past tense, and determines whether each of the plurality of prediction word candidates having a past tense is a past tense. The prediction input device, when the setting information of the past priority mode is stored in the tense priority mode storage memory, the prediction word candidate stored in the candidate output storage memory The prediction word candidates determined to be past tense are rearranged in a higher order and stored, and the prediction word candidates stored in the candidate output storage memory are displayed on the display device. The character input device according to claim 1. The tense term database stores a plurality of words representing the future as future priority candidates, and the tense priority mode storage memory stores future priority mode setting information for a sentence that includes words representing the future, If there is a future priority candidate corresponding to a search based on the tense term database using a word in a sentence created by the e-mail creation device as a medium, the predictive input device stores the future in the tense priority mode storage memory. Precedence mode setting information is saved, and the corresponding prediction word candidate is displayed on the display device based on the search of the prediction word database through the hiragana in the middle of input, and is selected by the e-mail creation device. A candidate for a predictive adjunct that is connected based on a search of the adjunct database for predicting connection using the predicted word as a medium When the setting information of the future priority mode is stored in the tense priority mode storage memory, the prediction type determined as the future form among the prediction auxiliary word candidates stored in the candidate output storage memory The candidate of an adjunct word is rearranged to a high rank, and is preserve | saved, The candidate of the adjunct word for prediction preserve | saved in the said memory for candidate output is displayed on the said display apparatus, The display apparatus of Claim 1 characterized by the above-mentioned. Character input device. In the character input method for predicting candidate adjunct words for connection for each word,
Searches for words in the text created for the tense term database that saves multiple past words as past priority candidates, and if there is a corresponding past priority candidate, the past priority is stored in the tense priority mode storage memory Saving the mode setting information;
A search word database that stores a plurality of prediction word candidates in the order of frequency of use through hiragana in the middle of input is searched using hiragana in the middle of input and the corresponding prediction word candidates are displayed on the display device. A process to display in the lower column of the text creation screen;
Prediction words are stored in order of frequency of use, and predictions are given to each connection prediction auxiliary word with information for determining whether it is a past tense. A search for a word for prediction selected by the display device with respect to a database for a word for use as a medium, and a corresponding candidate for a word for prediction to be saved is stored in a candidate output storage memory;
When the setting information of the past priority mode is stored in the tense priority mode storage memory, the prediction auxiliary word determined to be the past form among the prediction auxiliary word candidates stored in the candidate output storage memory Rearranging the candidates in a higher order and storing them, and displaying the prediction adjunct candidates stored in the candidate output storage memory in the lower column of the sentence creation screen of the display device;
Creating a text of an e-mail on the text creation screen of the display device by selecting a prediction word candidate or a connection prediction adjunct candidate displayed in the lower column of the text creation screen of the display device; A character input method comprising: In a character input program that predicts candidate adjunct words for connection for each word,
Searches for words in the text created for the tense term database that saves multiple past words as past priority candidates, and if there is a corresponding past priority candidate, the past priority is stored in the tense priority mode storage memory The procedure to save the mode setting information,
A search word database that stores a plurality of prediction word candidates in the order of frequency of use through hiragana in the middle of input is searched using hiragana in the middle of input and the corresponding prediction word candidates are displayed on the display device. The procedure to display in the lower column of the text creation screen,
Prediction words are stored in order of frequency of use, and predictions are given to each connection prediction auxiliary word with information for determining whether it is a past tense. A procedure for searching for an associated word database using a prediction word selected by the display device as a medium, and saving a corresponding candidate for an associated predictive word in a candidate output storage memory;
When the setting information of the past priority mode is stored in the tense priority mode storage memory, the prediction auxiliary word determined to be the past form among the prediction auxiliary word candidates stored in the candidate output storage memory A procedure for rearranging the candidates into higher ranks and storing them, and displaying the prediction adjunct candidates stored in the candidate output storage memory in the lower column of the sentence creation screen of the display device;
A procedure for creating an e-mail sentence on the sentence creation screen of the display device by selecting a prediction word candidate or a connection prediction candidate associated word displayed in the lower column of the display device sentence creation screen. A character input program comprising: In a mobile phone device that creates an e-mail by predicting candidates for attached words for prediction for each word,
The created text and the input hiragana are displayed on the text creation screen, and the prediction word candidates predicted to be prefetched or predicted adjunct words connected to the selected prediction word are displayed in the lower part of the text creation screen. A display unit to
A prediction word displayed on the display device, a prediction word selected on the basis of a connection prediction ancillary word candidate, an e-mail creation device that creates an e-mail sentence from the connection prediction ancillary word; and
Prediction word database that stores a plurality of prediction word candidates in order of frequency of use through hiragana in the middle of input,
Prediction words are stored in order of frequency of use, and predictions are given to each connection prediction auxiliary word with information for determining whether it is a past tense. An adjunct database for
A tense term database that stores multiple past words as past priority candidates,
A tense priority mode storage memory that stores past priority mode setting information for a sentence that includes a word representing the past and is being created;
A plurality of connection prediction ancillary word candidates passed from the connection prediction ancillary database are stored, and connection prediction word adjunct candidates that are determined to be a past tense with respect to a plurality of connection prediction ancillary word candidates are stored. Candidate output storage memory that is sorted and saved in the higher rank,
Setting information of the past priority mode in the tense priority mode storage memory when there is a corresponding past priority candidate based on the search of the tense term database using the words in the sentence created by the e-mail creation device The prediction word candidate corresponding to the prediction word database search based on the hiragana in the middle of input is displayed on the display device, and the prediction word selected by the e-mail creation device is displayed. Predictive adjunct word candidates connected to the connection prediction adjunct database search as a mediator are stored in the candidate output storage memory, and past priority mode setting information is stored in the tense priority mode storage memory. If there are prediction adjunct candidates stored in the candidate output storage memory, the prediction adjunct candidates determined as the past tense are ranked higher Rearranged and stored, the mobile phone apparatus comprising: a prediction input device for displaying the candidates of the candidate prediction appendage words stored in the output storage memory to the display device.

Images