JPS60110074A - Character processor - Google Patents

Abstract

PURPOSE:To limit the learning processing frequencies and to prevent the reduction of an input speed by performing the learning process to the homonyms immediately after conversion of KANJI (Chinese character). CONSTITUTION:A read string supplied from an input means 1 is converted into KANJI and phrases by a converting means 2, and the candidate KANJI for conversion (candidate words) are delivered to an output means 3. If the delivered KANJI (the 1st candidate word) is not equal to the desired KANJI and phrase, the next candidate word is converted by a next candidate indicating means 4 and the means 2 and delivered to an output means. When the delivered candidate word is coincident with the desired KANJI, etc., the selection is indicated by a homonym selecting means 5. Then the candidate words delivered presently to the means 3 are selected. In this case, it is decided whether the selected KANJI is equal to the one obtained immediately after the KANJI(Japanese syllabary)/KANJI conversion. If the selected KANJI is equal to the one obtained immediately after said conversion, the using frequency information or the using order information on a dictionary 7 is replaced by a learning means 6. Thus the learning process is performed to the homonyms immediately after the KANA/KANJI conversion. Therefore the input speed is never limited and the input efficiency is improved with the learning processing.

Description

DETAILED DESCRIPTION OF THE INVENTION [Technical Field] The present invention relates to a character processing device that edits and processes a Japanese document containing kanji and kana.

[Prior Art] In a device for inputting Japanese documents containing kanji and kana, when an operator inputs the reading of a kanji, a dictionary search is performed, a group of kanji candidates corresponding to the reading are displayed, and the candidates are selected. When a desired one from the group is selected by the operator, it is converted into the selected Kanji. Such a kana-kanji conversion input method has been widely adopted.

In this Kana-Kanji conversion input method, if the desired kanji is the first candidate word among the candidate words listed in the kanji dictionary, the desired kanji (first
Kanji conversion to candidate words) is performed. However, if the desired Kanji is a Kanji below the second candidate word, the desired Kanji must be selected using the next candidate key. As a result, the efficiency of kana-kanji conversion input increases as the probability that the operator's desired kanji is the first candidate word increases. Therefore, if a kanji is used even once by an operator, the usage frequency information or usage order information in the dictionary is updated as a frequently used kanji, and that kanji can be converted as the first candidate word when inputting the next kana-kanji conversion. So-called learning processing has been conventionally performed to improve the performance of the computer.

However, such learning processing requires manipulation of data in the dictionary as described above, so if the dictionary is stored on an external storage medium such as a floppy disk that takes time to access, the learning processing may not be executed. This has the disadvantage that input is prohibited during this period, which limits input speed.

In view of this, it is conceivable to not perform any learning processing at all, but in such a device, the kanji desired by the operator will not be converted as the first candidate word, so it is necessary to perform the candidate word selection operation. The disadvantage is that the number of times ``is increased'' and the input speed is slow.

[Purpose] The purpose of the present invention is to eliminate the above-mentioned conventional drawbacks, and to provide a character that takes advantage of the advantages of learning processing while limiting the number of processing times to prevent a decrease in input speed. The purpose of this invention is to provide a processing device.

[Example] The present invention will be described in detail below with reference to the drawings.

FIG. 1 shows the configuration of the present invention.

In the figure, a reading string inputted from an input means 1 is converted into kanji and compound words by a conversion means 2, and conversion candidate kanji (candidate words) are outputted to an output means 3. If the output kanji (first candidate word) is not the desired kanji or compound word, the next candidate instructing means 4 converts the next candidate word by the converting means 2 and outputs it to the output means. If the output candidate word is a desired kanji or the like, the homophone selection means 5 instructs selection, and the candidate word currently output to the output means 3 is selected. At this time, it is determined whether the selected kanji has just been converted to kana-kanji,
If the information has just been converted, the learning means 6 updates the usage frequency information or usage order information in the dictionary 7.

FIG. 2 shows the structure of an embodiment of the character processing device of the present invention.

In the illustrated configuration, the CPU is a microprocessor that performs calculations, logical judgments, etc. for character processing, and connects an address bus AB and a control bus CB.

Each component connected to these buses is controlled via the data bus DB. Address bus AB transfers address signals indicating the components to be controlled by the microprocessor CPU. The control bus CB transfers control signals for each component controlled by the microprocessor CPU. The data bus DB transfers data between each component device.

Next, the ROM is a read-only fixed memory, and the control procedure (9th
(Illustration transcription) etc. Further, the RAM is a writable random access memory having a configuration of 1 word and 18 bits, and is used for temporary storage of various data from each component. In this memory RAM, KBBUF is a keyboard buffer, which is a buffer for storing key data of the keyboard KB. C0BUF is a command buffer, which extracts key data from the keyboard buffer KBBUF one character at a time and temporarily stores it. The homophone buffer BUF indicates how many candidate words exist in the homophone. Further, MRCD is a buffer that stores a pointer to the most recently created homophone buffer BUF.

KB is a keyboard, which includes character symbol input keys such as alphabet keys, hiragana keys, and katakana keys, and various keys in this character processing device such as kanji part start/end indicator keys, conversion keys, next candidate keys, selection keys, etc. It is equipped with various function keys for instructing functions. DISK
is an external memory that stores documents created in the memory for storing fixed-form documents, and the stored documents can be recalled when necessary by instructions from the keyboard. A kanji dictionary is also developed in this memory DISK. C
R is a cursor register, and the contents of the cursor register can be read and written by the CPU. A CRT controller CRTC, which will be described later, displays a cursor at a position on the display device CRT corresponding to the address stored here. D
BUF is a buffer memory that stores document information etc. input from the keyboard KB. CRTC is a CRT controller and is responsible for displaying the contents stored in the cursor register CR and buffer [1BUF on the display CRT. Further, a CRT is a display device using a cathode ray tube or the like, and a dot-configured display pattern and a cursor display on the display device CRT are controlled by a CRT controller.

Furthermore, CG is a character generator, which generates character signals such as characters, symbols, and cursors to be displayed on the display device CRT.

This embodiment consisting of each of these components is a keyboard KB.
The microprocessor CPU operates in response to various inputs from the keyboard K13, and the input from the keyboard K13 is sent to the CPU, which reads various control procedures stored in the ROM and executes them. Various controls are performed according to the control procedures.

Figure 3 shows the structure of a kanji dictionary. As shown in the figure, the dictionary consists of 18 bytes per word.

The first 8 bytes are the reading, and each word has a reading of 1 character.
Stored in bytes. For example, the code is JISC knee 13
The lower byte of the H.228 code is used. The next 6 bytes are the kanji part, where each kanji character is stored in 2 bytes. The code is, for example, a JISC-622e code.The last two bytes are used to represent the frequency of use of the word. The usage order information is shown in the order in the dictionary, with the first word in the reading "Kikai" representing the usage order 1, and the second word representing the usage order 2.

Here, at the time of reverse character conversion, which of the homophones is to be selected as the first candidate word is determined by taking into account the degree of use and order of use of these words. In this embodiment, the candidate ranking is determined by the score calculated by C frequency of use - (1 order of use x 2)), and the word with the highest score of 9 is determined as the first candidate word.

FIG. 4 shows how the data in the dictionary changes due to the learning process.

Figure 4 (A) shows the initial state, and the highest score calculated in this state is 9 points for "machine", and when the reading "kikai" is input, "machine" is the first candidate word. It appears as.

FIG. 4(B) shows the state after selecting "R-meeting" from this state. As shown in the figure, the frequency of use of "Your Association" is counted up, and the order of use becomes the first. As a result, the score becomes the highest, and henceforth it can be converted as the first candidate word.

FIG. 4(C) further shows the state after "machine" is selected. As shown in the figure, the frequency of use of machine j is counted up, and the frequency of use becomes the first.

As a result, "machine" again has the highest score, and henceforth can be converted as the first candidate word.

FIG. 5 shows the structure of the document in the buffer memory DBUF. As shown, the document is divided into several lines, and each line is further divided into several characters.

FIG. 6 shows the structure of the character in more detail. As shown,
Each character consists of 2 bytes and is classified into two types depending on whether the first bit is 0' or '1'. In other words, as shown in FIG. 6(A), ``0'' indicates a normal character, and in each lower bit,
For example, character data such as hiragana, kanji, symbols, etc. is stored using the JISC-8228 code. However, as shown in Figure 6 (B), if the first bit is ``1pa'', it indicates that it is a homophone, and the lower 15 bits are sent to the homophone buffer, which indicates what kind of homophone it is. A pointer is stored.

FIG. 7 shows the structure of 1±homophone backer BUF. As shown in the figure, the homophone buffer BUF has a reading part that shows which pronunciation has been converted to become a homophone, a display number part that shows which candidate word is currently displayed, and the reading part of the reading part. It consists of a list of homophone candidate words corresponding to . The candidate word list is arranged in the order of candidates starting from the first candidate word. 1
A word consists of 6 bytes, a reading part consists of 8 bytes, a display number part consists of 2 bytes, and each candidate word consists of 6 bytes.

FIG. 8 shows the kanji conversion operation in this embodiment configured as described above.

FIG. 8(A) shows the screen of the display device CRT in a state where nothing is input. Here, S indicates a display area, and CM indicates a cursor. When characters are input using the keyboard KB, the characters are input at the position indicated by the cursor CM.

FIG. 8(B) shows the state from such a state.

This shows the state after inputting °° or ′°, ゛ii', ")'" and performing kana-kanji conversion. Here, "(" represents the kanji start key, and ")" represents the kanji end key. To convert kana to kanji, just surround the kanji reading with 1111 and ++)++ and then input the conversion key. In this case, there are multiple words (kanji) that correspond to the reading ``kikai'' in the dictionary, so the first candidate word ``machine'' is displayed as a homophone, circled. .

FIG. 8(C) shows a state in which the cursor CM is moved to the bottom of the screen from this state, and after inputting (naikaku), the conversion key is input to perform kanji conversion. Since there are multiple words with the pronunciation of ``Naikaku'' in the dictionary, the first candidate word ``Naikaku'' is displayed as a homophone. , FIG. 8(D) shows the state after pressing the next candidate key and then pressing the selection key in this state. The second candidate word for the reading of “Naika〈” is “cabinet”, so the kanji “
Cabinet" is confirmed. At this time, since "naikaku" has just been converted into kanji, the learning process is executed.

FIG. 8(E) shows the state after the next candidate key and selection key are pressed. The second candidate word for “Kikai” is “
Since the word is ``instrument'', the kanji ``instrument'' is confirmed. At this time, the selection of homophones for the reading ``kikai'' was not done immediately after the kanji conversion. In other words, it is considered that the word is not important enough for the operator to select it immediately after conversion, so no learning process is performed.

Next, FIG. 9 shows the operation of the character processing device according to this embodiment.

First, I, in step 9-1, the keyboard buffer K
It is detected whether key data is stored in BBUF, and if so, the key data is taken out to command buffer C0BUF and the following processing is performed. If the data is not stored, the process waits until valid data is stored in the buffer KBBUF. The data stored in the buffer KBBUF is created by an interrupt process that occurs when the second key of the keyboard KB is pressed.

In step 9-2, the key data in the buffer C0BUF is analyzed, and the process proceeds to subsequent steps according to the type of key. In other words, if it is a conversion key, proceed to step 8-3;
If it is a selection key, the process proceeds to step 8-5, and if it is any other key, the process proceeds to step 9-8. In step 9-8, various processes required for general character processing such as mode change and insertion processing are executed, and then the process returns to step l1l-1.

In step 8-3, kana-kanji conversion processing is performed.

Search the dictionary (see Figure 3) for the pronunciation sequence of the part surrounded by the kanji start code "(" and the kanji end code")" and convert it into kanji. At this time, if there is one candidate word in the dictionary, that candidate word is transferred as a conversion result to the buffer DBUF in which document data is stored. However, if there are multiple candidate words, all the candidate words are restored and the seventh
A homophone buffer BUF as shown in the figure is created, and a pointer pointing to the homophone buffer BUF is transferred to the document data in the /sisopha DBUF.

After this, in step 8-4, if there is one candidate word, 0"' is assigned to the buffer MRCD, and if it is a homophone,
That is, if there are multiple candidate words, a pointer for indicating the homophone buffer BUF of the M phonetic word is assigned to the buffer MRCD. In this way, buffer MRCD will have a pointer to the most recently created homophone buffer BUF. After this, step 9
Return to -1.

In step 8-5, the position of the cursor when the selection key was pressed is determined, and a search is made from the cursor position toward the beginning of the document to see if a homophone exists.

In this search, a character whose most significant bit on the document data is °'1 pa is detected. When the homophone search is completed, a pointer to the homophone buffer BUF is obtained from the code of the homophone in step 8-6, and the kanji code of the candidate word indicated by the display number section in the homophone buffer BUF is pointed to by the pointer. into the document. Next, the pointer to the homophone buffer requested in step 9-7 is compared with the value in the buffer MRCD, and if they do not match, it is not immediately after conversion, and the process returns to step 8-1. However, if they match, a learning process is performed in step 9-8, and then the process returns to step 8-1.

In the tree embodiment described above, the kana-kanji conversion is performed using the kanji specification method.

Of course, the present invention can also be applied to input using other methods, such as the clause specification method.

[Effects] As explained above, according to the present invention, since the learning process is performed on homophones immediately after kanji conversion, the dictionary does not exist on an external storage medium such as a floppy disk that takes time to access. Even in such cases, the input speed is not limited as in the past. In addition, for words that are used so frequently that input efficiency would be extremely low if they could not be converted as the first candidate word, the learning process is performed by selecting the candidate word immediately after converting the kanji. Since the candidate word becomes the first candidate word, it is possible to fully utilize the advantage of increasing input efficiency through learning processing.

[Brief explanation of drawings]

FIG. 1 is a diagram showing the overall configuration of the present invention, and FIG. 2 is a block diagram showing the configuration of an embodiment of the present invention. Figure 3 is a diagram showing the structure of a dictionary; Figures 4 (A) to (C) are diagrams showing changes in data in the dictionary due to learning processing; Figure 5 is a diagram showing the structure of a document; Figure 6 ( A) and (B) are diagrams showing the structure of characters, 7th
The figure shows the structure of the homophone buffer.
E) is a diagram explaining the operation of the apparatus shown in FIG. 2, and FIG. 9 is a flowchart showing the operation of the apparatus shown in FIG. DESCRIPTION OF SYMBOLS 1... Input means, 2... Conversion means, 3... Output means, 4... Next candidate instruction means, 5... Homophone selection means, 6... Learning means, 7... Dictionary, KBBUF, C0BUF, MRCD, BUF, DBUF
...Buffer, Patent Applicant Canon Co., Ltd. Fig. 3 θ height 6 height 2) (Item Fig. 4 Fig. 6 Fig. 7

Claims (1)

[Claims] It has a dictionary in which kanji are recorded together with their pronunciations, and multiple kanji having the same pronunciation are stored in a searchable manner in a predetermined order. Kanji conversion is performed by searching within the dictionary, and when there are multiple homophones for the input pronunciation in the dictionary, a desired homophone is selected from the homophone group, and the selection is performed using the kanji. A character processing device characterized in that a learning process is executed only when the learning process is performed immediately after conversion to change the searchable order in the dictionary of the homophone group to which the selected kanji belongs. . (Hereafter, margin)