JP2007004636A - Language input system, its processing method, recording medium and program - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a language input system improved in usability of user-related data. <P>SOLUTION: The language input system comprises a user dictionary file 203 storing a plurality of dictionary data 203a, 203b, 203c, 203d, and 203e by layer, the data set including a plurality of data records, and the data record including a data definition for acquiring a notation corresponding to reading. In the system, a sentence is analyzed, a data record including a data definition corresponding to the reading of a word contained in the sentence is extracted, and it is determined whether the extracted data record is contained in any one of the one or more data sets or not. When it is determined that the extracted data record is not contained in any one of the one or more data sets, a new data set containing the extracted data record is generated, and the generated new data set is stored in an empty layer of the storage means. <P>COPYRIGHT: (C)2007,JPO&INPIT

Description

  The present invention relates to a language input system, a processing method, a recording medium, and a program capable of displaying one or more notations for an input reading in a notation list and selecting and confirming one notation from the displayed notations.

  2. Description of the Related Art Conventionally, there is a language input system that operates a Japanese input program that converts an input language, for example, Japanese reading into desired Japanese notation (Hiragana, Katakana, Kanji, symbols, etc.). Examples of Japanese input programs include MS-IME (Input Method Editor) (trademark) of Microsoft Corporation, ATOK (trademark) of Just Systems Corporation, VJE (trademark) of Bucks Corporation, and others (see, for example, Patent Document 1).

  For example, MS-IME operates on a personal computer (PC) that operates in an OS (operating system) environment such as Windows (registered trademark). MS-IME refers to dictionary data (system dictionary) that MS-IME has in advance or dictionary data (user dictionary) defined by a user, and acquires one or more expressions for input readings from the dictionary. . MS-IME displays the notation as a notation list on the display screen of the PC, and converts the reading into one notation selected and fixed by the user.

  MS-IME is a program for inputting Japanese characters. MS-IME uses a lexicon database (DB) and a Japanese language model (LM) to generate appropriate ideograms (notation) from phonetic characters (reading). Lexicon is a glossary and a so-called dictionary of MS-IME. The LM is a grammar DB that defines rules for performing statistical language processing. The MS-IME extracts one or more notations corresponding to the input reading from the lexicon DB, and determines the optimum one notation by language processing using LM therefrom.

JP 2002-123510 A

  In the conventional language input system as described above, as seen in examples of abbreviations and new words born in the world, the vocabulary of Japanese changes every day, so the default DB of the Japanese input program is all Japanese Is not covered. Accordingly, the Japanese input program is required not only to correspond to a predetermined database but also to a user related DB that supports various words born in the world.

  The present invention has been made in view of such problems, and an object of the present invention is to provide a language input system, a processing method thereof, a recording medium, and a program that improve the ease of handling user-related data. It is in.

  In order to achieve such an object, the language input system of the present invention displays one or more notations for an input reading in a predetermined language, and displays one notation from the displayed notations. A language input system for selection, which is a storage means for storing one or more data sets (203a, 203b, 203c, 203d, 203e) for each layer, wherein the data set is one or more data Storage means (108) including a data definition for obtaining a notation corresponding to reading, and data including the data definition corresponding to reading of a word included in the sentence by analyzing the sentence Extraction means (101, S712) for extracting records, and the data records extracted by the extraction means include the one or more data sets The data record extracted by the extracted data determining means (101, S714) for determining whether or not it is included in any of the one or a plurality of data sets is included in the one or more data sets. If it is determined that there is no data generation unit (101, S716) that generates a new data set including the extracted data record and stores the generated new data set in an empty layer of the storage unit. , S718).

  In order to achieve the above object, the processing method of the language input system of the present invention displays one or a plurality of notations for an input reading in a predetermined language, and one of the displayed notations is displayed. A method of processing a language input system for selecting a notation, wherein the language input system is a storage unit that stores one or more data sets by layer, and the data set includes one or more data Including a record, the data record having a storage means including a data definition for obtaining a notation corresponding to the reading, and the method is such that the extracting means analyzes the sentence and corresponds to the reading of the word included in the sentence. An extraction step for extracting a data record including the data definition, and the data record extracted in the extraction step is included in the one or more data sets. In the extracted data determination step in which the extracted data determination means determines whether or not it is included, and in the extracted data determination step, the extracted data record is included in any of the one or a plurality of data sets. When it is determined that there is no data, the data generation unit generates a new data set including the extracted data record, and stores the generated new data set in an empty layer of the storage unit. And a step.

  In addition, the code | symbol etc. in the figure in embodiment corresponding to the component of a claim are shown by (). However, the constituent elements described in the claims are not limited to the constituent elements in the embodiment of the above () part.

  With the above configuration, the language input system can process various user-oriented vocabulary information by using a layered structure with a simple design. Moreover, the language input system can cover not only the well-known user situation corresponding to a predetermined DB but also a new category of vocabulary corresponding to a user-related DB that supports various words born in the world.

  ADVANTAGE OF THE INVENTION According to this invention, in a language input system, there exists an effect which improves the ease of handling of a user related data.

  Embodiments to which the present invention can be applied will be described below in detail with reference to the drawings. In addition, the same code | symbol is attached | subjected to the location which has the same function in each drawing, and duplication of description is abbreviate | omitted.

(Device configuration)
FIG. 1 is a system block diagram of a PC installed with a Japanese input program for realizing a Japanese input system for converting Japanese readings into desired Japanese notation in the present embodiment. In the present embodiment, a Japanese language input system using a general-purpose PC equipped with Windows (registered trademark) as an OS will be described.

  In the PC 100 of FIG. 1, reference numeral 101 in the figure denotes a CPU (central processing unit), which executes a program loaded in a random access memory (RAM) of the system memory 102. A system memory 102 is a RAM (Read Only Memory) that stores various data such as input data necessary for a program executed by the CPU 101, execution results of the program, and a BIOS (Basic Input / Output System). Etc. The RAM of the system memory 102 temporarily stores data to be displayed on the display 107 and data input from the keyboard 106, modem 103, microphone 117, and the like.

  Reference numeral 103 denotes a modem, which may be an internal modem or an external modem, and is connected to the system bus 116 via a serial port interface to establish communication on the wide area network 120 such as the Internet. Reference numeral 104 denotes a CD (compact disc) -ROM drive, which reads data from a mounted CD-ROM 105. In this embodiment, a program and data read from a CD-ROM 105 in which a Japanese input program and related data (system dictionary file or the like) are recorded are installed in a hard disk storage device (HD) 108 to be described later.

  Reference numeral 106 denotes a keyboard, and a character is input by pressing a key corresponding to the character. Reference numeral 117 denotes a microphone, which inputs the voice uttered by the user to the PC 100. Reference numeral 107 denotes a display that visually displays characters input from the keyboard 106 and the like and the calculation result of the CPU 101. The pointing device 115 can move a pointer (cursor figure) displayed on the display screen of the display 107 and can instruct the position for confirmation. In this embodiment, a mouse is used. . The pointer is moved by the user with the mouse 115 itself, and the position of the pointer is determined by the left and right click buttons. Reference numeral 116 denotes a system bus that connects the above-described elements of the PC 100.

  The HD 108 stores the following programs and data for the purpose of storage. Reference numeral 109 denotes an OS for controlling the PC 100 and peripheral devices. In the present embodiment, the above-described Windows (registered trademark) is used. Reference numeral 110 denotes various application programs other than the Japanese input program 114 of the present embodiment. For example, a browser for browsing a WEB page of WEB content, Explorer (trademark), word processor software, database software, spreadsheet Software, voice recognition software, etc.

  Reference numeral 111 denotes various program modules other than application programs, such as various device drivers and DLL (dynamic-link library) files that provide various APIs (Application Programming Interface). The program data 112 is various data used or generated by the OS and various other programs. The program data 112 includes various system dictionary files and user dictionary files associated with the Japanese input program of the present embodiment, and GUI (graphical user interface) data required for the Japanese input program to display a notation list. The system dictionary file supplied by the vendor includes system dictionary data, and the user dictionary file includes user dictionary data defined by the user.

  Reference numeral 114 denotes a Japanese input program, which is a program shown in the flowcharts of FIGS. 7 to 9 that defines the contents of the Japanese input program processing performed by the CPU 101. Details of this program will be described later in the following description of the operation.

  Reference numeral 120 denotes a wide area network, and the Internet is assumed in this embodiment. A WWW (World Wide Web) server 121 is connected to the wide area network 120. A web site of the vendor of the Japanese input program 114 that uses the system dictionary file of the present embodiment is published on the WWW server 121. In obtaining a new system dictionary file, it is possible to access the WEB site opened by the vendor of the Japanese input program 114 from the PC 100 via the network 120 and download the file from there.

  FIG. 2 is a system configuration diagram showing an outline of the operation of the Japanese input program 114. The Japanese input program 114 has a dictionary search module 114a and an LM module 114b.

  Basically, the dictionary search module 114a extracts one or more notations corresponding to the input reading 201 from the system dictionary file 202 in the vendor-supplied default lexicon DB and the user dictionary file 203 in the user lexicon DB. It is. Further, the dictionary search module 114a manages the update and deletion of data in the user dictionary file 203. The LM module 114b is a module that determines an optimal notation output 205 of one notation from one or more notations extracted by the dictionary search module 114a by language processing using the learning result file 204 in the LM.

  The system dictionary file 202 includes system dictionary data 202a. The user dictionary file 203 includes proper noun dictionary data 203a, E (electronic) mail response dictionary data 203b, abbreviation dictionary data 203c, document folder dictionary data 203d, new word dictionary data 203e, and the like. The learning result file 204 includes learning result data 204a.

  In particular, the present embodiment has a feature in the management method of each dictionary data in the user dictionary file 203 by the dictionary search module 114a. The management method of each dictionary data and each of these data in FIG. Detailed.

  The learning result data 204a records the learning result of the kana-kanji conversion process when the LM module determines the notation output 205. As a method of recording a learning result in LM, there are a word bigram method and various other methods. In this embodiment, a learning routine using a well-known flag count is applied.

(Description of operation)
In the system configuration described above, the operation of the Japanese input program 114 of the Japanese input system for converting Japanese readings of the present embodiment into a desired Japanese notation will be described below with reference to FIGS. .

  The Japanese input system of this embodiment processes various user-oriented vocabulary information by using a layered structure with a simple design. As a result, not only the known user situation corresponding to a predetermined lexicon DB but also a new category of vocabulary corresponding to the user lexicon DB that supports various words born in the world is covered.

  Further, the Japanese input system supports a user lexicon DB that can manage dictionaries of various categories, and the dictionary data of various categories uses a single format. The user-oriented vocabulary within each dictionary has a period-specificity. For example, a new word is generated by the public or the market, but sometimes such a new word is not used and the word is treated as a dead word. On the other hand, proper nouns such as personal names are permanently used by users. The Japanese input system is characterized in that it can easily handle the difference in the existence period between the user's dictionary data by using such layered data management.

  FIG. 3 shows, for example, when editing an e-mail response document by a mailer, the Japanese input program 114 refers to each dictionary file described above, and converts the input reading “None” into an appropriate notation obtained from each dictionary. It is a figure which shows the example in the case of doing.

  The Japanese input program 114 refers to each dictionary file and acquires one or more notations for the reading “None” input at the position of the reference numeral 301 in the drawing on the display screen. The Japanese input program 114 displays these notations as a notation list 302 on the display screen on the display 107 of the PC 100. At this time, the fraction “6/6” of the reference numeral 303 displayed in the notation list 302 has the notation “Hanaji” where the focus (highlighted for highlighting) 304 is located in all six notations. Indicates the sixth notation.

  The Japanese input program 114 moves the focus when the user depresses the down (down) key or the like of the PC keyboard 106, and closes the notation list 302 when the user presses the enter key (for example, the enter key). Then, the Japanese input program 114 converts the input reading “None” into a notation “Mr. Hana” selected and confirmed by moving the focus at the position of reference numeral 301.

  FIG. 4 is a diagram showing a method of managing each dictionary data in the user dictionary file 203 by the dictionary search module 114a. The user dictionary file 203 stores data in a layered structure. The user dictionary file 203 can store proper noun dictionary data 203a, e-mail response dictionary data 203b, abbreviation dictionary data 203c, document folder dictionary data 203d, new word dictionary data 203e, and the like as independent layer data. . FIG. 4 shows a case where the user dictionary file 203 already stores proper noun dictionary data 203a, abbreviation dictionary data 203c, and new word dictionary data 203e in each layer.

  When the e-mail response document 401 is started to be edited by the mailer, the dictionary search module 114 a analyzes the original message 401 a from the mail sender in the e-mail response document 401. The dictionary search module 114a extracts words in the original message 401a that are not present in each dictionary data existing in the system, creates e-mail response dictionary data 203b including the extracted words, and creates a free space in the user dictionary file 203. Store in the layer. At this time, the e-mail response dictionary data 203b is assigned an existing period “until mail transmission”.

  As a result, when the dictionary search module 114 a detects the email transmission of the email response document 401, the dictionary search module 114 a deletes the email response dictionary data 203 b from the user dictionary file 203. The layer in which the e-mail response dictionary data 203b is stored is an empty layer. It is determined that the e-mail response dictionary data 203b is unnecessary at the time when the words in the e-mail response dictionary data 203b are not used and are treated as dead words due to the existence period “until e-mail transmission” (when the e-mail response document 401 is transmitted). And can be discarded automatically.

  On the other hand, when the editing of the document file 402b stored in the folder 402 in the HD 108 is started by the word processing software or the like, the dictionary search module 114a analyzes the text of one or more other document files 402a in the folder 402. To do. The dictionary search module 114 a extracts words that are not included in each dictionary data in the text of the document file 402 a, creates document folder dictionary data 203 d including the extracted words, and stores it in an empty layer in the user dictionary file 203. . At this time, a predetermined existence period “until YYMMDD” is assigned to the document folder dictionary data 203d. In the expression format YYMMDD, YY represents the end of the Christian era (00 to 99), MM represents the month (01 to 12), and DD represents the day (01 to 31).

  As a result, when the dictionary search module 114 a detects that YYMMDD is exceeded, the document folder dictionary data 203 d is deleted from the user dictionary file 203. The layer in which the document folder dictionary data 203d is stored is an empty layer. Here, for the predetermined existence period “until YYMMDD”, an appropriate time preset in advance by the system is set assuming a future time when a word in the document folder dictionary data 203d is not used and is expected to be treated as a dead word. Just do it. As a result, the document folder dictionary data 203d can be automatically discarded when it becomes unnecessary.

  Further, various categories of dictionary data in the user dictionary file 203 use a single format throughout the entire dictionary. The dictionary search module 114a can easily handle the difference in the existence period between the dictionary data in the user dictionary file 203 by using such layered data management and a single format.

  FIG. 5 is a diagram for explaining a difference in existence period between dictionary data in the user dictionary file 203. In FIG. 5, the right direction shows the passage of time to the future. For example, a new word is generated by the public or the market, but sometimes such a new word is not used and the word is treated as a dead word. Therefore, the existence period 501c of the new word dictionary data 203e is defined by assigning a predetermined existence period “until YYMMDD” to the new word dictionary data 203e. Further, when the email transmission of the email response document 401 is completed, the words in the email response dictionary data 203b are not used and are treated as dead words. Therefore, the existence period 501b of the e-mail response dictionary data 203b is defined by assigning the existence period “until mail transmission” to the e-mail response dictionary data 203b.

  On the other hand, proper nouns such as personal names are permanently used by users. Therefore, the permanent existence period 501a of the proper noun dictionary data 203a is determined by assigning the existence period “permanent” to the proper noun dictionary data 203a. Since the system dictionary data 202a is data included in the system dictionary file 202 in the default lexicon DB supplied by the vendor, the existence period 501x of the system dictionary data 202a is permanent.

  FIG. 6 is a diagram showing the structure of the system dictionary file 202, user dictionary file 203, and learning result file 204.

  FIG. 6 shows a notation list table 601 having one or more notations for the reading “none”, which is acquired by referring to the learning result file 204 when the Japanese input program 114 displays the above-described notation list 302. The learning result file 204 is a file that stores learning result data 204a including one or more notation list tables corresponding to each reading.

  In FIG. 6, the notation list table 601 includes a plurality of records each having a reading field 602 (data is “none”), a notation field 603, and a flag count field 604. For example, in the record 605, the notation for the reading “None” is “Talk” and the flag count is “4”. The flag count indicates the priority order displayed at the top of the notation list 302 for the notation having the flag count. The notation possessed by the records in the descending order of the flag count of the records in the notation list table 601 is displayed in the notation list 302.

  Here, for example, at the position of reference numeral 301 in FIG. 3, it is assumed that the input reading “None” has been converted into one notation “spoken” selected and confirmed by the user moving the focus. In this case, the Japanese language input program 114 increments (++) the flag count of the record 605 in FIG. 6 to “5” by the LM module, and records the learning result of the kana-kanji conversion process when it is confirmed.

  FIG. 6 also shows a system dictionary table 606 having one or more notations for each reading. The system dictionary file 202 is a file in which system dictionary data 202a including such a system dictionary table 606, layer ID (blank), existence period (permanent), and data name (system dictionary data) is stored.

  In FIG. 6, the system dictionary table 606 includes a plurality of records each having a reading field and a notation field. For example, the record group 607 is read in order from the top, and the notation for “no, no, no, no, no, no” is “speak, talk, release, jealous, release”. Note that the record group to which the flag count is added to the record group 607 already exists in the notation list table 601 of the reading “None” in the learning result data 204a as the previous learning result.

  FIG. 6 also shows a proper noun dictionary table 608 having one or more notations for each reading. The user dictionary file 203 is a file in which proper noun dictionary data 203a including such a proper noun dictionary table 608, layer ID (1), existence period (permanent), and data name (proper noun dictionary data) is stored. FIG. 6 further shows a new word dictionary table 609 having one or more notations for each reading. The user dictionary file 203 is a file in which new word dictionary data 203e including such a new word dictionary table 609, layer ID (5), existence period (up to YYMMDD), and data name (new word dictionary data) is stored.

  Here, the case where the dictionary search module 114a described with reference to FIG. 4 creates the e-mail response dictionary data 203b and stores it in the empty layer in the user dictionary file 203 will be described with reference to FIG. In this case, if the layer with layer ID 2 (hereinafter referred to as layer 2) is an empty layer, the e-mail response dictionary data 203b is inserted and stored in layer 2.

  FIG. 6 shows an email response dictionary table 610 having one or more notations for each reading when email response dictionary data 203b is stored. The user dictionary file 203 stores e-mail response dictionary data 203b including such an e-mail response dictionary table 610, a layer ID (2), an existence period (until mail transmission), and a data name (e-mail response dictionary data). It becomes a file.

  In FIG. 6, the e-mail response dictionary table 610 includes a plurality of records each having a reading field and a notation field. For example, in the record 611, the notation for the reading “None” is “Hana”. Here, in synchronization with the insertion of the e-mail response dictionary data 203b into layer 2, a record in which the flag count 0 is added to the record 611 (reading “none”, notation “Hana”) is included in the learning result data 204a. The reading “None” is added to the notation list table 601.

  Next, the case where the dictionary search module 114a described with reference to FIG. 4 deletes the e-mail response dictionary data 203b from the user dictionary file 203 will be described with reference to FIG. In this case, when the e-mail response dictionary data 203b is deleted from the layer 2, the layers other than the layer ID of the layer 2 are blank and the layer 2 becomes an empty layer.

  Here, in synchronization with the deletion of the e-mail response dictionary data 203b from the layer 2, the above-described record in which the flag count is added to the record 611 (reading “none”, the notation “Hana”) is included in the learning result data 204a. Is deleted from the notation list table 601 of “None”.

  Similarly, in synchronism with the insertion of the email response dictionary data 203b into layer 2, records other than the record 611 (reading “None”, notation “Hana” ”) (for example, reading“ Tokio ”, notation“ Tokio ”) ) Is added to the notation list table of the reading “Tokio” in the learning result data 204a. Also, in synchronization with the deletion of the e-mail response dictionary data 203b from the layer 2, records other than the record 611 (reading “None”, notation “Hana” ”) (for example, reading“ Tokio ”, notation“ Tokio ”) In addition, the above-mentioned record to which the flag count is added is deleted from the notation list table of the reading “Tokio” in the learning result data 204a.

(Program processing explanation)
The processing related to the Japanese input in the present embodiment is performed according to the processing procedure shown in the flowcharts of FIGS. 7 to 9 show the contents of the Japanese input program processing executed by the CPU 101. The CPU 101 loads the Japanese input program 114 stored in the HD 108 into the RAM of the system memory 102 and executes it. Is done. Hereinafter, the processing procedure of the Japanese input program 114 performed by the CPU 101 will be described with reference to FIGS.

  Here, it is assumed that after the OS 109 is booted, the Japanese input program 114 is automatically resident in the system memory 102 and started by the OS 109. It is also assumed that an application such as a mailer or word processor software has been started in advance.

  In FIG. 7, the CPU 101 determines whether there is an event from the OS (step S700). When detecting an event from the OS, the CPU 101 determines whether or not the event is a document editing start event by the application (step S702). In the case of starting editing of a document, the CPU 101 inquires of the OS by a function such as an API, and acquires information as to which application is editing the document (steps S704 → S706).

  When the document is being edited by the mailer, the CPU 101 determines whether or not the editing of the e-mail response document 401 as shown in FIG. 4 is started (steps S708 → S710). If it is not the start of editing the e-mail response document 401, the CPU 101 performs the current processing of the mailer and returns to waiting for an event from the OS (steps S760 → S700).

  When the editing of the e-mail response document 401 is started, the CPU 101 analyzes the original message 401a from the mail sender in the e-mail response document 401, and records that have the reading and notation of the words included in the original message 401a ( (Hereinafter referred to as a word record) is extracted (step S712). Here, as an analysis method for extracting the word record, a known n-gram method, morphological analysis, or the like may be applied. Then, the CPU 101 stores each dictionary data existing in the system in the original message 401a among the extracted word records (system dictionary data 202a, proper noun dictionary data 203a, abbreviation dictionary data 203c, new word dictionary data 203e, etc.). New word records that are not present are determined, and those new word records are extracted (step S714).

  Next, the CPU 101 creates e-mail response dictionary data 203b including the new word record extracted in step S714 on the RAM of the system memory 102 (step S716). Here, as shown in FIG. 6, the e-mail response dictionary data 203b includes a layer ID (for example, 2) of an empty layer, an existing period (until mail transmission), a data name (e-mail response dictionary data), and a new word extracted. It consists of an e-mail response dictionary table 610 having records.

  The CPU 101 inserts and stores the e-mail response dictionary data 203b into an empty layer (for example, layer 2) in the user dictionary file 203 (step S718). Further, in synchronization with the insertion of the email response dictionary data 203b into layer 2, the CPU 101 adds a record related to the email response dictionary data 203b to the learning result data 204a (step S720). That is, as shown in FIG. 6, for example, the record 611 (reading “None”, notation “Hana” ”) and other records (eg, reading“ Tokio ”, notation“ Tokio ”) Each record to which 0 is added is added to each notation list table such as “None” or “Tokio” in the learning result data 204a. Then, the CPU 101 returns to waiting for an event from the OS (step S720 → S700).

  On the other hand, when the document is being edited by other than the mailer, the CPU 101 performs processing corresponding to the application that is editing the document in accordance with the processing in steps S710 to S720 (step S708 → S750). For example, the CPU 101 starts editing one or more other document files in the folder 402 when editing of the document file 402b stored in the folder 402 in the HD 108 is started by word processing software as shown in FIG. The sentence 402a is analyzed. The CPU 101 extracts words that are not included in each dictionary data in the text of the document file 402 a, creates document folder dictionary data 203 d including the extracted words, and stores it in an empty layer in the user dictionary file 203. At this time, a predetermined existence period “until YYMMDD” is assigned to the document folder dictionary data 203d. Then, the CPU 101 returns to waiting for an event from the OS (step S750 → S700).

  If the event from the OS is other than the start of document editing, the CPU 101 determines whether the event is a reading input by the keyboard 106 (steps S702 → S790). If the event from the OS is a reading input, the CPU 101 proceeds to the process shown in FIG. 8 (step S790 → FIG. 8). In cases other than reading input from the OS, the CPU 101 determines whether the event is a document editing end event by the application (steps S790 → S795). If the event from the OS is the end of document editing, the CPU 101 proceeds to the process shown in FIG. 9 (step S795 → FIG. 9). If the event from the OS is other than the end of document editing, the CPU 101 performs processing corresponding to the event and returns to waiting for an event from the OS (steps S795 → S700).

  In FIG. 8, when the event from the OS is a reading input, the CPU 101 stores information on the reading “nothing” input by the keyboard 106 as described above with reference to FIG. (Step S800). Next, the CPU 101 recognizes the notation list table 601 corresponding to the input reading “None” from the learning result data 204a (step S802). The CPU 101 edits in the buffer the data of the notation list 302 in which the notations of the records are arranged in descending order of the flag counts of the records in the notation list table 601 (step S804).

  Subsequently, when the CPU 101 detects an input event of the user's conversion key, the first notation (spoken) acquired from the data of the notation list 302 edited on the buffer in step S804 is used as the first conversion character string, and the reference numeral 301 indicates. The input reading “None” is displayed at the position (step S806).

  After displaying the first conversion character string, when the CPU 101 detects an input event of the confirmation key, an application (for example, a mailer) that is editing a document with the first conversion character string as the confirmation character string in a series of processes of steps S808 → S816 → S818. ) To perform post-processing (S816 → S818). The post-processing will be described later.

  After detecting the conversion key input event again after the first conversion character string, the CPU 101 displays the data of the notation list 302 edited on the buffer as the notation list 302 on the display screen of the display 107 as shown in FIG. Step S808 → S810).

  As shown in FIG. 3, the CPU 101 moves the focus 304 indicating the notation in the displayed notation list 302 downward to the input event of the conversion key and the down (↓) key, and inputs the up (↑) key. Move upward for an event. When the focus 304 exceeds the upper and lower limits of the notation list 302, the CPU 101 scrolls and updates the notation list 302. This process is performed by a loop between step S1812 and step S814. The CPU 101 always displays the notation instructed by the focus 304 as a converted character string at the position of the reading “None” inputted with reference numeral 301.

  When the CPU 101 detects a determination key input event in step S814, the CPU 101 performs post-processing (steps S816 → S818). As a post-processing, the CPU 101 first closes the notation list 302 in a series of processes from step S816 to S818, and passes the converted character string (the focused notation) to the application (for example, mailer) that is editing the document as a confirmed character string. .

  Here, for example, at the position of reference numeral 301 in FIG. 3, it is assumed that the input reading “None” has been converted into one notation “spoken” selected and confirmed by the user moving the focus. In this case, the CPU 101 increments (++) the flag count to “5” for the record 605 in the notation list table 601 of the learning result data 204a in FIG. Is recorded (step S820). Thereafter, the CPU 101 returns to waiting for an event from the OS (step S700 in FIG. 7).

  In FIG. 9, when the event from the OS is the end of document editing, the CPU 101 acquires the current time event from the OS and stores the information in a buffer on the RAM of the system memory 102 (step S900). As the current time event, information such as “mail transmission complete” and “YYMMDD” can be acquired as appropriate.

  Next, the CPU 101 stores the existing period in each existing dictionary data (such as proper noun dictionary data 203a, e-mail response dictionary data 203b, abbreviation dictionary data 203c, document folder dictionary data 203d, new word dictionary data 203e) in the user dictionary file 203. The acquired current time event is compared (step S902). For example, if the current time event is “mail transmission complete”, the CPU 101 determines that the e-mail response dictionary data 203b whose existence period is “until mail transmission” as shown in FIG. The process proceeds to step S904.

  In step S904, for example, the CPU 101 deletes the e-mail response dictionary data 203b in FIG. 6 whose expiration period has expired from the layer 2, and sets the layer 2 as a blank layer except for the layer ID of the layer 2 as a blank. Further, in synchronization with the deletion of the email response dictionary data 203b from the layer 2, the CPU 101 deletes the record related to the email response dictionary data 203b from the learning result data 204a (step S906). That is, as shown in FIG. 6, for example, record 611 (reading “None”, notation “Hana” ”) and other records (eg, reading“ Tokio ”, notation“ Tokio ”) are flagged for them. Each added record is deleted from each notation list table such as “None” or “Tokio” in the learning result data 204a. Thereafter, the CPU 101 returns to waiting for an event from the OS (step S700 in FIG. 7).

  On the other hand, if it is determined in step S902 that all existing dictionary data in the user dictionary file 203 is within the existing period, the CPU 101 returns to waiting for an event from the OS (step S902 → step S700 in FIG. 7).

(Effect of embodiment)
As described above, according to the present embodiment, as a first aspect, one or more notations for an input reading in a predetermined language are displayed, and one notation is displayed from the displayed notations. The language input system for selecting is a storage means that stores one or more data sets (203a, 203b, 203c, 203d, 203e) by layer, and the data set stores one or more data records. The data record includes a storage means (108) including a data definition for obtaining a notation corresponding to reading, and a data record including the data definition corresponding to reading of a word included in the sentence after analyzing the sentence The extracting means (101, S712) for extracting and the data record extracted by the extracting means are either one of the one or plural data sets. If the extracted data record is not included in any of the one or more data sets by the extracted data determination means (101, S714) for determining whether or not the data record is included Is determined, data generation means (101, S716, S718) for generating a new data set including the extracted data record and storing the generated new data set in an empty layer of the storage means. ).

  Here, as a second aspect, in the language input system according to the first aspect, each of the one or more data sets has a common format (FIGS. 4, 6, 203a, 203b, 203c, 203d). 203e), the data generation means may generate the new data set having the format.

  As a third aspect, in the language input system according to the first or second aspect, the data record included in the one or more data sets stored in the storage means corresponds to the reading and the reading. The data record extracted by the extracting means includes a reading of a word included in the sentence and a notation corresponding to the reading (610). .

  As a fourth aspect, in the language input system according to any one of the first to third aspects, each of the one or more data sets stored in the storage means has an existence period (see FIG. 6, 203a, 203e), and the data generation means assigns an existing period to the new data set (FIG. 6, 203b, S716).

  Further, as a fifth aspect, in the language input system of the fourth aspect, period determining means (101, S902) for determining each existing period of one or a plurality of data sets stored in the storage means; A data erasure unit (101, S904) that deletes from the storage unit the data set whose existence period has been determined to have expired by the period determination unit, and sets the layer in which the deleted data set is stored as an empty layer; Furthermore, it can be characterized by being provided.

  Further, as a sixth aspect, in the language input system according to any one of the first to fifth aspects, the sentence is an original message (401a) from the e-mail sender in the e-mail response document. Can be a feature.

  As a seventh aspect, in the language input system according to any one of the first to fifth aspects, the sentence is a content of a document file (402a) in a specific folder. it can.

  As an eighth aspect, in the language input system according to any one of the first to seventh aspects, the predetermined language may be Japanese.

  With the above configuration, the language input system can process various user-oriented vocabulary information by using a layered structure with a simple design. Moreover, the language input system can cover not only the well-known user situation corresponding to a predetermined DB but also a new category of vocabulary corresponding to a user-related DB that supports various words born in the world.

  In addition, since various categories of user-related data use a single format, the language input system can easily handle lifetime differences between user-related data by using such layered data management. it can.

(Other embodiments)
In addition to the embodiments described above, the following embodiments can be implemented.
1) In the above-described embodiment, for convenience of illustrating one embodiment, a plurality of dictionary data (203a, 203b, 203c, 203d, 203e in one user dictionary file 203 are shown in FIGS. Etc.). However, the present embodiment is not limited to this, and each dictionary data (203a, 203b, 203c, 203d, 203e, etc.) constitutes one user dictionary file and has a layer structure of the plurality of user dictionary files. It will be readily understood by those skilled in the art that variations are possible.

  Furthermore, although the case where one system dictionary file 202 includes one system dictionary data 202a has been described in the above-described embodiment, the present embodiment is not limited to this, and the system dictionary file includes a plurality of embodiments, or Those skilled in the art will readily understand that one system dictionary file can be modified into an embodiment including a plurality of system dictionary data.

2) In the above-described embodiment, an example in which kana-kanji conversion processing is performed on the reading “nothing” input at the position of the reference numeral 301 on the display screen of the display 107 input by the user using the keyboard 106. Explained. However, the present invention is not limited to kana-kanji conversion, but can be applied to all language processing systems. For example, the input of reading is not limited to a keyboard, and the present invention can also be applied to the case where a voice uttered by a user is input to the PC 100 with a microphone and the result of the recognition of the voice by voice recognition software is captured and processed. Will be easily understood by those skilled in the art.

3) In the above-described embodiment, an example of a Japanese input system that processes Japanese is shown, but the present invention is not limited to Japanese. Those skilled in the art will readily understand that the present invention is applicable to various languages having homonyms.

4) In the above-described embodiment, an example in which a Japanese input system is realized by a general-purpose personal computer has been described. However, as an information processing device other than a personal computer, a workstation, a server, a PDA (Personal Digital Assistant), a mobile phone, The present invention can be applied to other information processing apparatuses capable of executing various programs.

5) The recording medium referred to in the present invention refers to a medium in which a program executed by the CPU is recorded and can be read by the device. As the recording medium, in addition to the CD-ROM, a known recording medium such as an IC (integrated circuits) memory, an HD, a floppy (registered trademark) disk, a magneto-optical disk (MO) can be used.
The program recorded on the recording medium may be any of the program itself, a compressed program, and an encrypted program, and all these data are included in the concept of the program of the present invention.
Further, when transferring (downloading) a program to an information processing apparatus via a network such as the Internet or a LAN (local area network) or a signal line, a recording medium or storage device for storing the program of the transfer source apparatus is provided. This corresponds to the recording medium of the present invention.

6) Although the above-described embodiment has been described for the purpose of illustrating the present invention, modifications can be made in addition to the above-described embodiment. As long as the modification is based on the technical idea of the present invention described in the claims, the modification is within the technical scope of the present invention.

It is a system block diagram of PC which installed the Japanese input program of embodiment which can apply this invention. It is a system configuration | structure figure which shows the operation | movement outline | summary of the Japanese input program of embodiment which can apply this invention. It is a figure which shows the example in the case of converting into the input appropriate reading notation of embodiment which can apply this invention. It is a figure which shows the management method of each dictionary data in the user dictionary file by the dictionary search module of embodiment which can apply this invention. It is a figure explaining the difference in the existence period between the dictionary data in the user dictionary file of embodiment which can apply this invention. It is a figure which shows the structure of the system dictionary file of embodiment which can apply this invention, a user dictionary file, and a learning result file. It is a flowchart which shows the Japanese input program process which CPU of embodiment which can apply this invention performs. It is a flowchart which shows the Japanese input program process which CPU of embodiment which can apply this invention performs. It is a flowchart which shows the Japanese input program process which CPU of embodiment which can apply this invention performs.

Explanation of symbols

100 Personal computer 101 CPU
102 System memory 103 Modem 104 CD-ROM drive 105 CD-ROM
106 Keyboard 107 Display 108 HD
109 OS
110 Various application programs 111 Various program modules 112 Program data 114 Japanese input program 115 Mouse 116 System bus 117 Microphone 120 Wide area network 121 WWW server 114 Dictionary search module a
114b LM module 201 Input reading 202 System dictionary file 202a System dictionary data 203 User dictionary file 203a Proper noun dictionary data 203b E (electronic) mail response dictionary data 203c Abbreviation dictionary data 203d Document folder dictionary data 203e New word dictionary data 204 Learning result File 204a Learning result data 205 Notation output 401 E-mail response document 401a Original message 402 Folder 402a, 402b Document file 501a, 501b, 501c, 501x Existence period 501a

Claims (18)

A language input system for displaying one or more notations for an input reading in a predetermined language and selecting one notation from the displayed notations,
Storage means for storing one or more data sets by layer, wherein the data set includes one or more data records, and the data records include a data definition for obtaining a notation corresponding to reading. Storage means;
Extracting means for analyzing a sentence and extracting the data record including the data definition corresponding to the reading of the word included in the sentence;
Extracted data determination means for determining whether the data record extracted by the extraction means is included in any of the one or more data sets;
If the extracted data determination means determines that the extracted data record is not included in any of the one or more data sets, a new data set including the extracted data record is obtained. A language input system comprising: data generation means for generating and storing the generated new data set in an empty layer of the storage means. Each of the one or more data sets has a common format;
The language input system according to claim 1, wherein the data generation unit generates the new data set having the format. The data record included in the one or more data sets stored in the storage means includes a reading and a notation corresponding to the reading;
The language input system according to claim 1, wherein the data record extracted by the extraction unit includes a reading of a word included in the sentence and a notation corresponding to the reading. Each of the one or more data sets stored in the storage means has a lifetime.
The language input system according to any one of claims 1 to 3, wherein the data generation unit assigns an existing period to the new data set. Period determination means for determining the existence period of each of the one or more data sets stored in the storage means;
A data erasure unit that deletes the data set whose existence period is determined to be expired by the period determination unit from the storage unit, and sets the layer in which the deleted data set is stored as an empty layer. The language input system according to claim 4, wherein: The language input system according to claim 1, wherein the sentence is an original message from an e-mail sender in an e-mail response document. The language input system according to claim 1, wherein the sentence is a content of a document file in a specific folder. The language input system according to claim 1, wherein the predetermined language is Japanese. A language input system processing method for displaying one or more notations for an input reading in a predetermined language and selecting one notation from the displayed notations,
The language input system is storage means for storing one or more data sets for each layer, and the data set includes one or more data records, and the data records obtain a notation corresponding to reading. Storage means including a data definition for the method,
An extracting step of extracting the data record including the data definition corresponding to the reading of the word included in the sentence by analyzing the sentence;
An extracted data determination step in which an extracted data determination means determines whether the data record extracted in the extraction step is included in any of the one or more data sets;
In the extracted data determination step, when it is determined that the extracted data record is not included in any of the one or more data sets, a data generation unit includes the extracted data record A language generation system processing method comprising: a data generation step of generating a new data set and storing the generated new data set in an empty layer of the storage means. Each of the one or more data sets has a common format;
The processing method of the language input system according to claim 9, wherein in the data generation step, the new data set having the format is generated. The data record included in the one or more data sets stored in the storage means includes a reading and a notation corresponding to the reading;
The processing method of the language input system according to claim 9 or 10, wherein the data record extracted in the extraction step includes a reading of a word included in the sentence and a notation corresponding to the reading. Each of the one or more data sets stored in the storage means has a lifetime.
The language input system processing method according to claim 9, wherein an existing period is assigned to the new data set in the data generation step. A period determination step in which the period determination unit determines the existence period of each of the one or more data sets stored in the storage unit;
A data erasing step, wherein the data erasing step deletes the data set whose existence period is determined to have expired in the period determining step from the storage means, and sets the layer in which the deleted data set is stored as an empty layer; The language input system processing method according to claim 12, further comprising: The processing method of the language input system according to claim 9, wherein the sentence is an original message from an e-mail sender in an e-mail response document. The processing method of the language input system according to claim 9, wherein the sentence is a content of a document file in a specific folder. The processing method of the language input system according to claim 9, wherein the predetermined language is Japanese.   17. A computer-readable recording medium on which a program for causing a computer to execute each step of the processing method of the language input system according to claim 9 is recorded. A program for causing a computer to execute each step of the processing method of the language input system according to any one of claims 9 to 16.

Images