JPS60105029A - Document input device - Google Patents

Abstract

PURPOSE:To decrease the number of input man-hours of a user by producing automatically the phrase spelling supplied previously just with input of a part of the head spelling of the phrase in an estimated control mode. CONSTITUTION:In an estimated control mode, a reading circuit 21 is started by a signal line L13 and a character string on a screen buffer 18 corresponding to a cursor is set to an input character buffer 22. At the same time, a comparators 23 is started. The comparator 23 compares the contents of the buffer 22 with the head of a character string 26 of a self-contained word table 11. This table 11 stores the string 26 of phrases fed so far and its generating frequency 27. As a result of this comparison, each coincident entry of the table 11 is set to a candidate table B24. At the same time, a writing circuit 25 is started to read each entry of the table B24. Thus a character string of the highest generating frequency is extracted and set at the position of a cursor work register 10 on the buffer 18 via a data bus L18.

Description

[Detailed Description of the Invention] [Field of Application of the Invention] The present invention relates to a device that transforms Japanese sentences input in Kanase or Roman letters into Kanji or mixed sentences, and in particular, the invention relates to a device that converts Japanese sentences input in Kanase or Romaji into Kanji or mixed sentences. Relates to a device for inferring specific idioms from the first part of the phrase.

[Background of the invention]

The mainstream Japanese input method is the kana-kanji conversion method, in which the user inputs kana or romaji and the system automatically converts them into corresponding kanji. A problem with creating documents using this kana-kanji-hensei method is that the same compound word often occurs many times in one document, and therefore the ``gate'' is also the same. I often had to go through the hassle of inputting the Japanese kana spelling. especially,
This is rarely a problem for ordinary users who don't remember the layout of kana or alphabet letters on the keyboard. During this time, IJAlr: In order to solve the problem, I used a macro input function. When using this war story, users (
i. Register the result of kana-kanji conversion 7 in advance to a specific key ruru or macro name, and when creating a document, use the registered key ruru (ha) as the macro name t- as necessary. Then, the corresponding phrase was inserted into sentence-H. However, even with this method, which AM was registered?
Also, I forgot which key word I had registered the idiom in the macro name, and it turned out that I didn't use it very often. Also, I found it troublesome to have to register each idiom one by one.

[Purpose of the invention]

The purpose of the present invention is to improve such conventional problems, so that when a user wants to input the same idiom again as a previously inputted idiom, the user only needs to input part of the ``gishi'' at the beginning of the idiom. An object of the present invention is to provide a document input device that automatically generates spellings of previously inputted phrases. By the way, user O
The purpose is to reduce the number of incoming crabs.

[Summary of the invention]

The document input device of the present invention operates in a predictive control mode by automatically storing kana or romanized compound words (excluding adjuncts) and their frequency of occurrence that are input by the user when creating a document. In this case, each time a kana or romaji character is input, the input 4 words are compared with the spelling words of the idioms already written, and the most frequently occurring one is displayed from among the matches. control. With this control, the user can input the complete spelling of an idiom once, and thereafter only need to input a part of the spelling at the beginning of the idiom. If the expected idiom is not displayed, just enter the next spelling of the idiom. As a result, another idiom will be newly displayed on the screen. You can move the cursor to Yang 1, where the expected compound word is displayed, and then input an auxiliary word or another compound word.

In the normal mode, display control of the phrase is not performed, and the user inputs a special character added to a part of the spelling at the beginning of the phrase. As for the arrangement, for "Menpine Seigyosoute", inputs are made such as "Tafupine 9" and "Yu/%" (abbreviation input). The document input device detects the input 7 as a character string, and compares the previous spelling υ with an already occurring idiom to find a match. Perform 1fftl paralysis, which converts everything into kanji. 2. The predictive control mode is activated by pressing a specific key. Hereinafter, a concrete explanation will be given using one implementation sequence.

[Embodiment of the Invention] An embodiment of the present invention will be explained with reference to FIG. Code l in the diagram
is the keyboard for inputting kana and romaji money; 2 is the conversion key for instructing kana-kanji conversion; 3 is the predictive control mode key for starting the predictive control mode; 4 is for keeping the cursor on the front screen. A register 5 is a register that holds a character code input from the keyboard; 6 is an input character code determination circuit that determines the contents of the input character code register 5 and activates each control circuit; and 7, predictive control n
Register to be sent when mode key 3 is pressed, 8
is a line key l1tIl# circuit that performs processing for keys such as kana and romaji, 9 is an AND circuit, and 10 is a register that holds the start position of the input character string (in the case of writing from bunsetsu, the start of each bunsetsu remember the position), 1
Reference numeral 1 indicates an independent word table IIVc that stores the kana or romaji spellings of idioms input by the user and their frequency of occurrence, and 12 records the kana or romaji origin of the idioms input by the user. 13 is an input character prediction control circuit that compares the ibs of idioms and displays the one with the highest occurrence frequency among the idioms; 14 is a kana-kanji conversion control circuit, and 15 is a circuit that sets the input abbreviation in the case of an abbreviation input (the arrangement is a cous inputted as "tanmatsu%" for "tanman seigyeonkuchi"). 16 is a control circuit that compares the abbreviations with spellings of compound words stored in the independent word table 11; 17 is a table that sets all independent words that match the comparison; 18;
1 represents a screen buffer corresponding to the contents of the display screen 20 on an l-to-l basis; 19 represents a control circuit for displaying the contents of the screen buffer; and 20 represents the display screen of the terminal. Also, codes LX-L2, L7 to L10, L14 to L18, L25
~L27, L29~L32 are data buses, L3~L
6. Lll~L13. L28 represents a signal line.

Next, the flow of the operation shown in FIG. 1 will be explained. On the Kana/Romaji keyboard, the characters keyed in to convert to Kanji or the input codes for each key (conversion key 2, prediction side leg mode key 3, etc.) are input character codes via the data bus L1. - The current cursor position is stored in register 5 via data command bus L2e.
are set respectively. Furthermore, the input character code
The contents of register 5 and cursor register 4 are input to input character code determination circuit 6. Here, the contents of the input character code register 5 are determined and each control circuit is activated. If conversion key 2 is input,
When line 18 L3 is logic "1" K, the kana-kanji conversion control circuit 14 is activated.

When the prediction 1bu control mode key 3 is input, the signal line L4 becomes logic "1", and the prediction control start register 7
is set. When a blank character is input, the signal aL5 is set to logic "1" and the contents glass 1 of the cursor register 4 is set via the data control bus L9. The calf #-work register IO is a register that holds the starting position of a valid character string excluding blanks. The blank character is the delimiter between each clause. If characters other than the above are entered (kana,
For alpha bent characters, etc., the signal) line L6 is set to logic "1".
"Next, normal key ll+ljr+11 circuit 8 is activated. Each control wholesale circuit is connected to a data bus L7. The input character code and cursor address are respectively sent via L8. Normal key control circuit 8 sends the input character code to the cursor address position on screen buffer 18 via data bus LIO. The normal key system @1 circuit 8 further controls the logic of the signal and gland Lll.
Make it. Pre-jllJ control [When in Kl mode, believe @ plot L
The logic of 12 is j9 in shield Kr1J, and J in AND circuit 9:
Then, the 1 word line Lll and 1,120 are ANDed, and the logic of the signal line L13 becomes "1". Signal, 1lli
! When the logic of Lla becomes rlJ, the input character jllJ control circuit 12 is activated. When not in the prediction control sieve mode, the input character prediction control circuit 12 is not activated.

The operation of the input character prediction control circuit 12 will be explained using FIG. The cursor work register 10, cursor register 4, independent word table 11, and screen buffer 18 in the figure are the same as in FIG. 21 in the figure is a character string reading circuit from the screen buffer 718, and 22 is a circuit for reading character strings from the screen buffer 718.
A buffer 23 is used to set the read characters (input characters) (The content of the buffer 22 is compared with the kana string 26 of the independent word table 11, and the matching entry is set in the independent word table auxiliary table B24v). circuit, 25
represents a circuit that writes the character string with the highest frequency among the entries set in the independent word candidate table B24 to the screen buffer 18. Ma7hi, code L
14-L20. L22 to L23 are data buses, Li
2. L21°L24 represent signal lines. next,
The flow of the operation shown in FIG. 2 will be explained. The cursor work register lO indicates the start position of the input character string on the screen buffer 18 (in the case of writing from segments, the start position of each clause), and the cursor register 4 indicates the current cursor position on the screen buffer lB. keeping. In other words, the character string between the cursor work number register IO and the cursor battle register 40 on the screen buffer 18 indicates the character string currently input by the user (in the case of writing from bunsetsu, one sentence S). -ru. The readout circuit 21 from the screen buffer receives the signal I! It is started by JL13.

The signal line L13 is the output signal of the AND circuit 9 in FIG. The readout circuit 21 from the screen buffer is connected to the data bus L15. L14t-, the contents of the cursor work group register lO and cursor register 4 include R, and the character string present in the corresponding position of both on the screen buffer 18 is input to the input character buffer 22r. Set via bus L19. And signal J
The logic of tL2i is set to "1" and the comparison circuit 23 with the independent word table is activated. The comparison circuit 23 with the independent word table compares not only the contents of the input character buffer 22 but also the data.
At the same time, each entry in the independent word table 11C) is read in via the data bus J, 16, and a comparison is made between the two. The independent word table 11 stores character strings 26 in kana or Roman letters of phrases input by the user so far, and their frequency of occurrence 27. An independent word table comparison circuit 23 compares the contents of the input character buffer 22 with the beginning of the character string 26 of the independent word table 11, and transfers each entry of the independent word table 11 that matches the two to the data bus L22. via the independent word candidate table B24. Then, the logic of the signal rmL24 is set to "l" and the writing circuit 25 to the screen buffer is activated. The screen buffer writing circuit 25 reads each entry of the independent word candidate table B24cD via the data bus L23i. Extract the most frequently occurring character string from the read entries and transfer that string to the cursor work register lO position on the screen buffer 18 via the data bus L1.
8", cent via H. FIG. 3 shows an example of how the contents of the screen buffer 18 change as the input character control circuit 12 operates. shows that when the final input of the character string ``johoushori'' is only ``jo'', ``johoushori'' is automatically generated. 28 in the figure indicates the current cursor position. When the user first inputs "ji" at the current cursor position (18-2), the input character control circuit compares "ji" with the beginning of the character string 26 of the independent word table 110. ``Jikan'', ``Jouhou Shori''.

Matches each entry of "Sif/". Among these, "deer/" with the highest frequency of occurrence is set in the screen buffer (
is-a). At this time, the cursor 28 is positioned at the next character. The user then enters "Yo" because the expected character string is not displayed. The input character control circuit compares the character string "jo" input by the user so far with the beginning of the character string 26 in the independent word table 11. The character string "JOUHOSHORI" is written and the character string is set in the screen buffer (18-5). At 18-6, the expected character string was displayed, so the cursor was moved and a blank was input, indicating fc result. The following is an explanation of the input character prediction control circuit that performs the main control of the present invention.

Next, referring again to FIG. 1, when the conversion key 2 is input, the kana-kanji conversion control circuit 14 operates to change the logic of the signal line L3 to rlJ. The kana-kanji conversion control 14 extracts kana or roman characters on the screen buffer 18, converts them into kanji, and sets the conversion results back into the screen buffer 18 via the data bus L25. . When converting kana or romaji to kanji, the conversion dictionary 13 is used. Furthermore, the kana-kanji conversion control circuit 14 performs processing to create each entry of the independent word table 11, which is the main table of the present invention. Each entry consists of the kana or romaji number of the idiom and its frequency of occurrence.

When registering an idiom, it is determined whether or not the same 4-shi exists in the entry, and if it exists, the frequency of occurrence is increased by 1. If it does not exist, a process of adding a new IJ is performed. Data movement during this process is performed via data bus L26. The kana-kanji conversion control circuit 14 also performs cutting of abbreviations in the abbreviation input control of the present invention operating in the normal mode. When converting kana or romaji on the screen buffer 18 to kanji, it is detected whether special characters are added to the kana or romaji (or not), and in this case, the special characters are removed and the data bus is It is set in the abbreviation buffer 15 via L27 (for example, in the case of "you%", "you" is set). At this time, 'IN section L
28 to rlJ to activate and assist the abbreviation input state circuit 16. Abbreviation human power f[flJ The leg circuit 16vi reads the contents of the abbreviation buffer 15 via the data bus L29, and at the same time reads each entry of the independent word table 11 via the data II bus L30, and creates a four-word buffer. 19 and the beginning of the character string in the independent word table 11, and all matching entries are set in the independent word candidate table A17.

The kana-kanji conversion control circuit 14 reads each entry of the independent word candidate table A17 via the data bus L32, and converts all character strings therein into kanji. The result of converting each character string into Kanji is controlled as a homophone. FIG. 4 shows a column in which the character string "JIYO%" on the screen pan 7a 18 is converted to "information processing" through the independent word table 11. The character string "JIYO" is extracted as an abbreviation and compared with the beginning of the character string 26 in the independent word table 11. - Only "Jouhou Shori" is selected as the attack, converted into kana-kanji, and "information processing" is set in the screen buffer 18.

Next, referring again to FIG. 1, the contents of the screen buffer 18 are finally displayed on the display screen 20 through the display control circuit 19i.

〔Effect of the invention〕

As explained above, according to the present invention, when creating a document using kana-kanji conversion input, it is possible to input the complete 4 shi of an idiom in kana or romaji once, and from then on, for the same idiom, You only need to enter part of the first four syllables of the idiom. Generally speaking, many identical idioms occur in a single literary class, so it is possible to reduce the number of mistakes.

[Brief explanation of the drawing]

FIG. 1 is a block diagram of a document input device, FIG. 2 is a block diagram of an input character prediction control circuit, FIG. 3 is a diagram showing a specific sequence of input control, and FIG. It is a figure showing the concrete syllabary by inputting an abbreviation.

Claims (1)

[Claims] 1. Equipped with kana character keys or alphanumeric keys, converts kana character input or multi-romanization input into kanji or 1,000 name names.
Shadow? ta - I am at the specified document input device, and I
A special key called a control mode key, a means for recording the kana or romaji spelling of an idiom input in the past and its spelling 90 occurrence frequency, and a means for recording the spelling and the currently input spelling. The control circuit to be compared and detects whether the kana, ruru, or romaji spelling input to the comparison control circuit matches a part of the beginning of an idiom that has already been written down.
A document input device comprising a control means for displaying a matched binding if a match is detected.