JPH03282875A - Document preparation device - Google Patents

Abstract

PURPOSE:To decrease the number of candidates and to reduce the burden at the time of selecting a target word by eliminating the word whose use frequency is low from in the same sound word candidate. CONSTITUTION:A CPU 11 extracts a word corresponding to input reading information as the same sound word candidate from in a long-term learning dictionary 31, and rearranges this extracted same sound word candidate in order of high use frequency in accordance with frequency information. Subsequently, when the rearrangement of words is finished, the CPU 11 checks whether a word whose frequency information reaches a prescribed level H by long-term learning exists in its same sound word candidate or not, and detects the word having frequency information being below a level L set in advance as a word whose use frequency is low. Next, the CPU 11 eliminate the word whose use frequency is low, detected in such a way from in the same sound word candidate. Accordingly, a target word is selected from in the candidates limited to some extent, and the burden at the time of selecting the same sound word is reduced.

Description

DETAILED DESCRIPTION OF THE INVENTION [Object of the Invention (Field of Industrial Application) The present invention relates to a document creation device having a long-term word learning function.

(Prior Art) Conventionally, in a document creation device, when performing homophone conversion, among a plurality of words with the same pronunciation, words can be referred to in order of frequency of use.

This is achieved by using a dictionary that stores a plurality of words with the same reading and frequency information for each word, and sequentially reading words corresponding to the input reading information from this dictionary according to the frequency information.

By the way, in a document creation device that uses such a dictionary, if the frequency information of words is fixed, even if the word is particularly used by the user, if the frequency information of that word is low, it may not appear immediately. This may cause problems such as not being able to move. Therefore, a learning function is generally used that updates the frequency information of the selected word each time the user selects one of a plurality of words for the same reading. This allows the dictionary to have a word appearance frequency structure tailored to the user.

This learning function includes long-term learning and short-term learning.

In long-term learning, unless there is a specific instruction, frequency information is updated according to the frequency of word appearance. On the other hand, in short-term learning,
For example, the frequency information is initialized each time the power is turned on or each document. Here, a document creation device equipped with a long-term learning function will be described.

Conventionally, all words with the same reading were learned over a long period of time, and these words were displayed as homophone candidates during homophone conversion. In this case, the candidates include words that are rarely used and are rarely used by the user. Normally, homophone conversion is performed by the user selecting a target word (word to be converted) from among displayed candidates. However, if the candidates include words that are rarely used, it becomes difficult to select the desired word.

(Problem to be solved by the invention) As mentioned above, in the past, homophone candidates included even words that were used infrequently, resulting in a large number of candidates, making it difficult to select the desired word. There were other problems.

The present invention was created in view of the above points, and is a long-term solution that reduces the number of candidates by removing infrequently used words from homophone candidates, thereby reducing the burden of selecting a target word. The purpose is to provide a document creation device equipped with a learning function.

[Configuration of the Invention (Means for Solving the Problems) That is, the present invention has a storage means that stores a plurality of words for the same pronunciation and frequency information of each of the words, and the frequency is determined according to the frequency of appearance of the word. In a document creation device equipped with a long-term learning function that updates information, homophone candidates corresponding to the input reading information are extracted from the storage means, and frequency information among the homophone candidates reaches a predetermined level through long-term learning. If there are words that are
The system displays candidates with that word removed.

(Operation) According to the above configuration, words that are rarely used and are rarely used by the user can be removed from the homophone candidates. Therefore, by referring to the candidates, the desired word can be easily selected.

(Embodiment) Hereinafter, a document creation device according to an embodiment of the present invention will be described with reference to the drawings.

FIG. 1 is a diagram showing the configuration of a document creation device according to the same embodiment. In the figure, 11 is a microprocessor (CP
U), 12 is a memory accessed by the microprocessor 11. Microprocessor (hereinafter referred to as CP)
Reference numeral 11 controls the entire device, accessing the memory 12 and starting programs that follow input instructions, including document creation processing, long-term learning processing, etc.
Candidate control processing as shown in FIG. 3 or 4 is executed.

The memory 12 includes a program area 13 in which various processing programs that determine the operation of the CPU II are stored.
A character pattern area 14 in which font data is stored, a dictionary area 15 in which various dictionaries for converting readings input in kana or Roman characters into kanji are registered, and a frequency level area 25 are provided.

Here, in the same embodiment, the dictionary area 15 of the memory 12
A long-term learning dictionary 31 as shown in FIG. 2 is stored. This long-term learning dictionary 31 has a word appearance frequency structure due to the long-term learning function, and the dictionary 31 contains a plurality of words with the same reading and the number of times each word has been used up to now. In other words, frequency information is stored. Further, the frequency level area 25 stores various types of information necessary when removing words with low usage frequency from homophone candidates.

In addition, 16 is a keyboard (KB), 17 is a keyboard interface (KB-IF), 18 is a display device,
19 is a display controller (display-CNT), 20 is a display dot memory, and 21 is a floppy disk device (FD).
D), 22 is an FDD interface (FDD-IF
), 23 is a printer device (PRT), and 24 is a printer interface (PRT-IF).

The keyboard (KB) 16 uses the keyboard interface (KB-IF) '17 to input information necessary for document creation, etc.
Output to CPU II via. This keyboard 1B includes "character" keys, "numeric" keys, "cursor" keys, etc., as well as a "select/execute" key for instructing confirmation of various settings, etc. Various function keys are provided, including a "delete" key, an "insert" key, etc.

The display device 18 is composed of, for example, a 10-line liquid crystal display, and the bottom line (10th line) is set as an area exclusively for guide display (guide line). Display-CNT19 is C
Under the control of the PUII, the display of the display 18 is controlled using the display dot memory 20.

The floppy disk device (FDD) 21 is a CPU II
under the control of the FDD interface (FDD-IF
) 22 to save the created document or use external characters (J
Characters and symbols other than IS 1st level characters and 2nd level characters
marks, etc.) and second-level characters, etc.

The printer device (PRT) 23 is connected to the printer interface (PRT-IF) under the control of the cPU 11.
24, documents stored in the document area in RAMII or document information read into the main body from the floppy disk device (FDD) 21 and stored in the document area are printed on paper, or the paper is transported. Execute print processing such as

Next, the operation of this embodiment will be explained.

As shown in the flowchart of FIG. 3, when reading information is input through K B 1B (step AI), the CPU
1l extracts the word corresponding to the manual reading information from the long-term learning dictionary 31 as the same language candidate. Step A2)
, the extracted homophone candidates are sorted in descending order of frequency of use according to the frequency information (step A3).

To explain concretely, for example, if you input the pronunciation "Kosei", as shown in Figure 2, "Composition", "
welfare”, “rehabilitation”, “proofreading”, “offensive”, “fairness”, “
The words ``Kosai'', ``Hagane'', and ``Posterity'' are extracted as homophone candidates. In this case, the frequency information of "configuration" is highest in rloOJ. Therefore, this "structure" comes first, followed by "welfare,""offensive," and "rehabilitation."

Sorting is done in the order of "proofreading", "posterity", "fairness", "made of steel", and "Kosai".

In this way, when the word sorting is finished, the CPU
II first checks whether there is a word whose frequency information has reached a predetermined level H through long-term learning among the homophone candidates (step A4). Here, if the frequency information has reached the predetermined level H, it can be determined that there is a certain degree of difference in the frequency of use of each word. In other words, it can be determined that among the homophone candidates, words that are frequently used are distinguished from words that are rarely used.

When the CPU II confirms a word whose frequency information has reached a predetermined level H among the homophone candidates, the CPU II detects a word having frequency information below a preset level L as a word with low usage frequency (step A5). And C.P.
The UII removes the infrequently used words thus detected from the homophone candidates (step A6) and displays the candidates (step S7). Note that the level H and the level L are stored in advance in the frequency level area 25 of the memory 12.
is set to any value.

As a result, in the example of FIG. 2, for example, level H is "100".
'', and if level L is set to ``10'', the first place will be ``composition'' for the reading ``niusei''.
will be displayed, and the following will be displayed: "Welfare", "Offensive", "Rehabilitation", "
"Proofreading" and "Posterity" will be displayed in this order. In this case, "Fair", "Steel", and "Kosai" are level "1".
0" or less, it is not displayed as a candidate. In this way, infrequently used words that are rarely used by the user are excluded from candidates.

Therefore, the user can select a target word from candidates that are limited to a certain extent.

In addition, when homophones including words excluded from homophone candidates are selected through such candidate control processing,
For example, it is easily possible to provide a candidate-group display function and display all candidates related to the frequency of long-term learning by pressing a certain key.Also, for affix words and simple clauses, If this candidate control process is not performed, such as by attaching a flag to the word corresponding to this, the conversion rate of homophones can be maintained without decreasing.

Further, in the above embodiment, a reference level L for words that are used less frequently is set in advance, and words that are used less frequently are detected by comparison with this level L, but the present invention is not limited to this. Instead, for example, as shown in FIG. 4, words that are used less frequently may be detected based on their frequency difference with words that are used more frequently. The operation will be explained below.

That is, as shown in the flowchart of FIG. 4, when homophone candidates for the pronunciation are obtained (step 8
1-B3), CPUII found among these homophone candidates,
When the existence of a word whose frequency information exceeds a predetermined level H is confirmed (step B4), the difference between the word's frequency information (if there are multiple words, the highest frequency information among them) is determined in advance. Words with a set level M or higher are detected as words with low usage frequency (step B5). Note that level M is set in memory 1 in advance, similar to the above level HSL.
The frequency level area 25 of No. 2 is set to an arbitrary value.

To explain more specifically, in the example shown in Figure 2, if level H is rloooJ and level M is set to "90", the frequency difference with "composition", which is a word of level H or higher, is level The words "fairness", "steelwork", and "kosai", which have M or more words, are detected as words that are used infrequently.

Thereafter, as in the case of FIG. 3, the CPU II removes the infrequently used words thus detected from the candidates (step B6) and displays the candidates (step B7).

In this way, even if words with low usage frequency are detected and removed based on the frequency difference, the same effect as in the above embodiment can be obtained. In this case, when the long-term learning level (frequency information) of a certain word reaches level H, only the frequency information of that word is increased (for example, by 3 times) to increase the frequency difference, thereby increasing the Words may be excluded from the candidates.

[Effects of the Invention] As described above, according to the present invention, words with low usage frequency are removed from homophone candidates. Therefore, a target word can be selected from candidates that are limited to a certain extent.

This can reduce the burden of selecting homophones.

[Brief explanation of the drawing]

FIG. 1 is a block diagram showing the configuration of a document creation device according to an embodiment of the present invention, FIG. 2 is a diagram showing the configuration of a long-term learning dictionary of the embodiment, and FIG. 3 is a candidate control process in the embodiment. Flowchart for explaining operation FIG. 4 is a flowchart for explaining candidate control processing operation in another embodiment of the present invention. 11... Microprocessor (CPU), 12...
Memory, 15... Dictionary area, 16... Keyboard (
KB), 17...Keyboard interface (KB)
-IF), 18...Display device, 19...Display controller (display-CNT), 20...Display dot memory, 31...Long-term learning dictionary.

Claims (3)

[Claims] (1) In a document creation device that has a storage means that stores a plurality of words corresponding to the same pronunciation and frequency information of each word, and is equipped with a long-term learning function that updates the frequency information according to the frequency of appearance of the word, an input pronunciation is provided. candidate extraction means for extracting homophone candidates corresponding to the information from the storage means; and whether there are words whose frequency information has reached a predetermined level through long-term learning among the homophone candidates extracted by the candidate extraction means. a word detection means for detecting a word with low usage frequency from among the homophone candidates if there is a word whose frequency information has reached a predetermined level as a result of the frequency check means; A document creation device comprising: candidate control means for removing the word detected by the word detection means from the homophone candidates; and display means for displaying the homophone candidates obtained by the candidate control means. (2) In a document creation device that has a storage means that stores a plurality of words corresponding to the same pronunciation and frequency information of each of the words, and is equipped with a long-term learning function that updates the frequency information according to the frequency of appearance of the word, the input pronunciation is candidate extraction means for extracting homophone candidates corresponding to the information from the storage means; and whether there are words whose frequency information has reached a predetermined level through long-term learning among the homophone candidates extracted by the candidate extraction means. a frequency checking means for checking whether or not the frequency information of the word has reached a predetermined level as a result of the frequency checking means; word detection means for detecting a word as a word with low usage frequency; candidate control means for removing the word detected by the word detection means from the homophone candidates; and displaying the homophone candidates obtained by the candidate control means. 1. A document creation device comprising: display means. (3) In a document creation device that has a storage means that stores a plurality of words corresponding to the same pronunciation and frequency information of each word, and is equipped with a long-term learning function that updates the frequency information according to the frequency of appearance of the word, the input pronunciation is candidate extraction means for extracting homophone candidates corresponding to the information from the storage means; and whether there are words whose frequency information has reached a predetermined level through long-term learning among the homophone candidates extracted by the candidate extraction means. and a frequency check means for checking whether the frequency information reaches a predetermined level as a result of the frequency check means, and if there is a word whose frequency information has reached a predetermined level, the frequency difference between the word and the word whose frequency information has reached the predetermined level is determined from among the homophone candidates. a word detection means for detecting a word having a preset level or higher as a word with low frequency of use; a candidate control means for removing the word detected by the word detection means from the homophone candidates; 1. A document creation device comprising: display means for displaying obtained homophone candidates.