JPS621062A - Documentation supporting device - Google Patents

Abstract

PURPOSE:To obtain easily a unified document by checking the level distribution showing the spoken words or written words to the words of the whole document and unifying those words of both types to either type having a larger number of words. CONSTITUTION:A documentation part 2 performs the KANA (Japanese syllabary)/ KANJI (Chinese character) conversion and stores the produced sentences in a sentence storage part 3. An analyzing part 4 analyzes the sentence structure and extracts words. A comprehending part 5 retrieves a dictionary part 6 and stores the words themselves, the end-form, the parts of speech and the grammar information as well as the level value in a word storage part 7. For the word levels, the spoken words and the written words are set at 1 and -1 respectively with other words set at 0 for example. The part 5 compares the level mean value with the threshold value and stores them in a level mean value storage part 8 after normalization to 1 or -1. A correcting part 9 compares the word level with the level mean value to send them directly to a display part 10 as long as no difference exists between them. If the codes are opposite to each other, the part 6 extracts the end-form of the adverse level. Then the conjugation is produced from the parts of speech and the grammar information extracted out of the part 7 and sent to the part 10.

Description

DETAILED DESCRIPTION OF THE INVENTION [Technical Field of the Invention] The present invention relates to a support device such as a word processor or a workstation.

[Masumei's technical overview and its problems]

When a human composes a document, for example,
Spoken words and U-shaped words are irregular (-). This also happens with the usage of ``Ojigana'' and the distinction between ``Kiji'' and ``Hiragana''. If you use a document created by someone and take things too lightly, you will end up with an inconsistent document.

For example, when saying ``I bought......I bought...'', it would be more polite to say either one of them.

However, in Jubai Karijutsu 2, humans found inconsistencies and corrected them. Especially when there is a large amount of Chapter X, this task becomes a heavy burden for humans.

[Purpose of the invention]

SUMMARY OF THE INVENTION An object of the present invention is to provide a support device for a machine to assist a human in the creation of a unified and structured text.

[Summary of the invention]

The present invention reads sentences that have already been written manually, analyzes the sentence structure, and extracts words. Search the dictionary for that word,
Get the level of that word. Level is a number that indicates the difference between spoken language and written language.

Then, the levels are obtained for word 1 in the entire document, the distribution of the obtained levels is evenly distributed, and the words to be changed are limited in order to unify the distribution toward the higher level. The words thus limited are searched in the dictionary and converted into words of the desired level.

〔Effect of the invention〕

According to the present invention, it is possible to easily obtain a uniform document, which greatly improves the practical advantages for the user when creating a document.

[Implementation sequence of the invention]

Hereinafter, one embodiment of the present invention will be described with reference to the drawings.

Figure 1 is a schematic configuration diagram of the implementation/examination device, Figures 2, 3,
FIG. 4 is a control flow diagram thereof.

The input unit 1 is, for example, a txD from a keyboard input device.

It is configured with a plurality of character code keys, various control code keys, etc. The above character codes and various control codes are passed to the text creation section 2 t:C. (FIG. 2, 11) The sentence creation section 2 performs kana-kanji conversion based on the kana character code and control code passed from the input section 1, and stores the resulting sentence in the sentence storage section 3.

(Fig. 12) The analysis unit 4 performs syntactic analysis on the sentence read from the sentence storage unit 3, extracts words, and passes the word, its final form, part of speech, and grammatical information to the understanding unit 5. (Fig. 3, 13.14) The understanding part 5 searches the dictionary part 6 for the final form of the word passed by the analysis part 4, and obtains the level of that word. (Fig. 3, 15) In this embodiment, the level takes the value 1 if it is spoken Wx, -1 if it is written, and O if it is neither.In addition, the dictionary section 6 contains the level and this value for each car #. Words corresponding to the opposite level are stored respectively.The understanding unit 5 takes in all the words passed to the analysis unit 4 and takes the average value of the level.Then, the obtained average value is determined in advance. If it is smaller than a certain threshold value, it is understood that the input sentence is mainly written in written language, and conversely, if it is larger than the threshold value, it is understood that the input sentence is mainly written in spoken language. For all the words obtained here, the understanding section 5 stores the word itself, its final form, part of speech, grammatical information, and its level value in the word storage section 7 (Fig. 3, 16). 5 is a comparison of the obtained average level value with the threshold value 1, and the average value of the level is normalized to 1 or -I C2 depending on either the spoken words or the written words that are 1 folded on 9. and store it in the level average value storage section 8. Then, the calibration section 9
Request the next stage of processing. (Fig. 3 17) Calibration section 9
Now, the sentence is called up again from the starting position from the sentence storage section 3, and it is broken down into word units through the analysis section 4 (Fig. 4, 18).
At the same time, the words, their final forms, parts of speech, grammatical information, and their levels generated by the understanding unit 5 are sequentially read out from the word storage unit 7 (see FIG. Compare the level of the issued word with the level average value. If the level is the same as the value stored in the level average value storage section, it is sent to the display section 10 as is. (Fig. 4 21.25) On the other hand, if the signs are reversed, the dictionary unit 6 retrieves the final form of the opposite level of the word (Fig. 4 22), and the final form and the word storage unit 7 A conjugated form is generated using the extracted part of speech and grammatical information and sent to the display section 10. (Fig. 4 23
．． 24) At this time, a decoration is displayed with high brightness to notify humans that the level has been changed.

In the proofreading section 9, single chains are sequentially extracted from the sentence storage section 3 and the word storage section 7, and the level is changed repeatedly until the screen of the display section 10 is filled. (Fig. 26 of 弗4) Therefore, the level can be returned to the original level by asking for an instruction from the input unit 1 and by a human instruction. (Figure 4 2
7) In that case, search the word storage unit 7 for the word displayed at the position indicated by the person, obtain its final form, part of speech, and grammatical information, and then retrieve the final form and the level. Search the dictionary section 6 using the inverted value with the sign as the level,
The final form is obtained, and a conjugated form is generated using the final form, the part of speech and grammatical information retrieved from the word storage section 7, and the conjugated form is sent to the clothing section 10 to be rewritten, and at the same time, it is sent to the sentence storage section 3. p
Change its contents. This correction work is repeated for the word corresponding to the changed level displayed on the screen, and a human instructs the next screen to select the same level, completing the text storage section 30P3'#.

Therefore, using this device eliminates the difficulty of reading the input text due to the mixture of profanity and profanity, and there is still some human intervention required. However, since the machine shows the points to be corrected, it is now possible to create documents with great ease of use.

Note that the present invention is not limited to the above embodiments. For example, the understanding section 5 can assign a level to the word C2 read out from the sentence storage section without human intervention.
However, the level obtained at this time may be sent to the display s1o along with the words and confirmed interactively.

Furthermore, what is inputted to the input unit 1 may be character codes from the keyboard (position coordinates as handwritten information, etc.), or a system may be used in which a handwritten position fW@ is recognized to obtain a kanji code.

Furthermore, instead of searching the dictionary section 6 based on the final form of the word, the salt S section 5 may search the word using its inflected form and convert it into a form that can be searched in the dictionary section, if necessary.

stomach.

In addition, we adopted either spoken or written language as the level of the word, but it is possible to choose between the writing style of ``de-aru tone'' and ``1-masu tone'', and the style in which it is said that it is okay to use ``nome-9'' or omit it. You may also write it in kanji or use hiragana.

Furthermore, although the understanding unit 5 compares it with a certain threshold value, this threshold value may be changed externally.

Further, the level average value storage section 8 may not be provided.

In that case, the proofreading unit 9 reads the sentence again from the starting position from the sentence storage unit 3, breaks it down into word units through analysis #4, and sends the obtained word to the understanding unit 5 to form its final form. , part of speech, and grammatical information (: All you have to do is compose.

Further, although the level is set to l or -l, the value may be an ordinary number.

In addition, when the original word was changed depending on the level of the word in the proofreading section 9, it was displayed in high brightness to notify humans. It may also be possible to change the color.

Furthermore, although the manual input to the proofreading section 9 has been described for the entire document, it may also be for a part of the document.

Also, in the display when the calibration part 9≦2, when the screen is filled, the power that makes it possible to have a conversation with a human (this means that if you wait for the screen to be filled, 1 spelling 1 first C 2 It may be configured to process.

Also, in the proofreading department 9, the words were changed conversationally twice, but
Functions such as changes may be included without conversation (- all at once). In short, the present invention can be implemented with various modifications without departing from its gist.

[Brief explanation of drawings]

FIG. 1 is a diagram showing one embodiment of the present invention, FIG. 2, and FIG.
FIG. 4 is a diagram showing a part of the control flow. l... Input section 2... Sentence creation section 3...
Textbook tii part 4...Analysis part 5...J! ! Solution part 6... Dictionary part 7... jlL word storage part
8...Level average position Vls9...Proofreading department
40... Ibarashibe agent Patent attorney Nori Chika
Yu (and 1 other person) Figure 1 Figure 2 Figure 3

Claims (1)

[Claims] A sentence storage unit for storing already created sentences, an analysis unit for analyzing the sentences stored in this sentence storage unit and extracting words, and a level and level indicator for each word to indicate whether it is spoken or written. A dictionary section stores words corresponding to this reverse level, and a level corresponding to each word extracted by the analysis section is searched from the dictionary section, and each of these words is stored in correspondence with its relative level. An understanding section that calculates the average value, and a word corresponding to the opposite level in the dictionary section based on the result of comparing the level of each word stored by this understanding section with the average value, thereby determining the meaning of the sentence. A document creation support device comprising: a proofreading section that proofreads each word into spoken or written words; and a display section that displays each word of the sentence proofread by the proofreading section.