JPS62200456A - Document processor - Google Patents

Abstract

PURPOSE:To draw out quickly a sentence at a later time by storing a key word annexed to its sentence at the time of registering the sentence in a sentence storage part. CONSTITUTION:A user inputs two sentences of 'the 90th anniversary of the foundation of our company' and ' memorial party is held', as sentence retrieval request sentences by using a sentence input part 1. As a result, a sentence analyzing part 7 extracts three words of plural words 'company', 'foundation', and 'year' from said sentence by a KANJI (Chinese character) dictionary 8, converts them to words (key words) of broader terms, and extracts three key words 'company, commencement, and time'. Subsequently, a sentence subject control part 9 retrieves a subject name containing its key word from a sentence relation table storage part 10, based on these three key words, and picks out an optimum subject name from in several subject names. Next, the sentence of the document file name and sentence number concerned is drawn out of a sentence storage part 4 and displayed in order of a candidate by a sentence output part 6.

Description

DETAILED DESCRIPTION OF THE INVENTION <Field of Industrial Application> The present invention relates to a document processing device that can efficiently create new documents by extracting document examples such as Japanese documents and English documents from a storage device.

<Prior art> Conventionally, when creating a relatively formal document such as a business document using a document processing device such as a Japanese word processor or a Roman word processor, a relatively formal document that has been created and registered in the past is used. It was often possible to create a new document efficiently by using a document almost as is or by only making some modifications.This was also a major advantage of word processors.

FIG. 2 shows a block diagram of a conventional document processing device.

In the figure, 1 is a character input section consisting of a keyboard, etc., and 2 is a character input section 1 for inputting identification codes, etc. classified according to the content of the text.
A text retrieval unit extracts the desired sentence (formal short sentence) from the text storage unit 4 such as a magnetic disk or optical disk by inputting the text from the text storage unit 4; 6 a document creation/correction unit for creating new sentences; 5 a text registration unit for registering relatively formal sentences in the text storage unit 4 in a format that is easy to use later; is a document output unit that displays the information required when performing each of the above processes on a CRT screen, etc. 0 When a user registers a text in the text storage unit 4 in the above conventional document processing devices, By attaching an identification code or the like to the card, later withdrawal can be facilitated. However, in order to attach this identification code to a document, it is essential to appropriately classify each document in advance. If this is not done, the number of registered documents will increase - the identification code will include many sentences, and it will take time to retrieve the most appropriate sentence, or it will be difficult to retrieve the appropriate sentence if the identification code is attached by mistake. The problem was that it was not possible.

<Purpose of the Invention> The present invention solves the above-mentioned conventional problems by easily and appropriately storing information (keywords) associated with a text when registering the text in the text storage unit.
It is an object of the present invention to provide a new text processing device capable of quickly extracting the text later.

<Embodiment> Hereinafter, an embodiment of the document processing apparatus according to the present invention will be described in detail with reference to the drawings.

FIG. 1 shows a block diagram of an embodiment of a document processing apparatus according to the present invention. In this figure, the same parts as in FIG. 2 are designated by the same reference numerals. In FIG. 1, 1 is a character input unit consisting of a keyboard, etc., 2 is a text search unit that retrieves the desired text from a text storage unit 4 such as a magnetic memory or optical memory, and 3 is a unit that corrects the text extracted from the text storage unit 4. Document creation/editing department for writing or creating completely new sentences, 6
displays one sentence of information required for each process on a CRT
This is a document output unit that displays on a screen, etc.

Reference numeral 7 denotes a sentence analysis unit, and when the user inputs a sentence outlining the request content (text search request sentence) through the character input unit 1, the sentence analysis unit 7 uses a dictionary 8 to perform processing based on the request sentence. A plurality of words (single in some cases) are extracted, and a superordinate concept word (keyword) is derived from the extracted word group. Reference numeral 9 denotes a text subject control unit, which narrows down the subject of the text intended by the user (to be created) from several subject names. Reference numeral 10 denotes a text relationship table storage unit, and the text relationship table storage unit 10 associates the above-mentioned keywords with sentences having the keywords.

A search for a desired sentence using the document processing device having the above configuration is executed as follows. Incidentally, FIG. 3 shows the text search algorithm.

For example, when a user creates a document based on the intention of ``I want to hold a 90th anniversary party,'' the procedure is as follows.

The user uses Bunjin Sakibu 1 as a text search request sentence, ■
Input the following two sentences: ``It's our company's 90th anniversary.'' ■ ``We're holding a commemorative party.'' (101). As a result, the sentence analysis unit 7 uses the kanji dictionary 8 shown in FIG.
``company'' = ``company'',
Convert to words (keywords) of higher-level concepts on the right side of the arrows, such as "founding" - "start" and "year" - "time", and extract the three keywords "company 1 start 1 hour" (102
).

Through the above processing, three keywords are extracted by the sentence analysis section 7.

Next, based on the above three keywords, the text subject control unit 9 searches the text relationship table storage unit 10 shown in FIG. 5 for a subject name (this is also used for the document file name) that includes the keyword (103); The most suitable subject name is narrowed down from several subject names (104).

It should be noted that sentences are stored in the sentence storage section 4, categorized by subject. The text search unit 2 refers to the text relationship table storage unit 10 based on the subject matter narrowed down by the text subject control unit 9, and searches for the document file name and text number that contain the most of the above keywords. Furthermore, all searched sentences are sorted in order of comprehensiveness (105). Then, the text with the corresponding document file name and text number is retrieved from the text storage unit 4 and displayed in the order of candidates in the document output unit 6 (10
6).

In this case, the subject name of the text presented from the text storage unit 4 is "Guidance" for the above request sentence (■), and the presented text is "
Now, thanks to your support, our company will soon be celebrating its founding anniversary. ”. The keywords included in this proposed sentence are "confidential, company, intimate 9 start 1 hour, pick up", and this keyword and the extracted 3
This sentence was selected because it had something in common with the keyword ``Company 2 start 1 hour''. At this point, the user recognizes that this presented text is what he was looking for and selects it (1
07), the device determines that the subject of the text that the user is currently creating is "guidance" (108).

Here, since there is a second request statement (109),
A search is performed for this request statement. As for the request sentence of above ■, the theme has already been determined as "guidance", so based on this theme, (110) the sentence relation table storage unit 10 is referred to and the request sentence of above ■ is derived. The keyword “recall,
Sentences with keywords matching "banquet, event" are presented. That is, the sentence analysis unit 7 extracts three words "memorial", "party", and "event" from the sentence (■) using the dictionary 8, and divides this word group into "memorial" - "reminiscence".
, "Party" - "Banquet", "Event" = "Event", etc. are converted into the words (keywords) of the higher-level concept on the right side of the arrow, and the three keywords "Recall."

Extract "banquets, ceremonies". Finally, the text output section 6 displays the text "In this regard, we will hold a heart-warming banquet." It should be noted that there is no problem even if the document search for the request statement (①) is performed independently from the case for the request statement (②), and the search is performed on a separate subject.

Next, the user modifies the text retrieved from the text storage section 4 using the document creation/modification section 3. In this way, the user creates a document while modifying the retrieved text.

Next, registration of one document is executed as follows.

In addition, FIG. 6 (((shows the algorithm for text registration).

When the user executes the procedure for registering the corrected text in the text storage unit 4 (201), the document processing device presents the subject name to which the original text used to create the text belongs (201).
202) L, with the user's consent (203), determines the subject name of the text (204). If the user's consent is not obtained, all subject names are presented and it is left to the user to decide which subject name to use among them (205). Once the subject name of the created text is determined (in this case, guidance), the text file name (annai) associated with the subject name is to be stored. Then, use a text number that has not yet been used in the text file name (normally, the text number of the new text is one greater than the maximum number n of existing text numbers in the text file name). Registration is performed and a text index (the index in the part (C) of FIG. 5) is created (206). Next, keywords included in the sentence are extracted by the sentence analysis section 7. That is, the sentence analysis unit 7 extracts words from the created sentence using the dictionary 8, and converts this word group into words (keywords) of a superordinate concept, thereby extracting a plurality of keywords. Then, for each of the extracted keywords, the subject, purpose 1 sentence number, and sentence formation (described in detail later) are determined.
Incorporate into the text relationship correspondence table shown in Figures (a) and (b) (2
07) o In addition, the classification of usage (internal, 2 external) is inherited from the original text. However, like the others, this usage name can be changed by the user. In this creation work, it is necessary to modify the existing sentence relationship correspondence table, and it is necessary to insert a sentence number for each keyword in the central block (b) of the sentence relationship correspondence table shown in FIG.

Here, the format of the above sentence (A in the sentence relationship correspondence table).

B, C+ free, etc. ) is a convenient encoding of the parts indicated by ``'' in the following sentences, for example.

In the above format, "free" is used when a text that does not fit into a normal format, and whose content is frequently used, is not included in the text relationship correspondence table.

Sentences registered using the above method are checked at the time of registration using a subject name that is very close to the gist of the text, so there are fewer registration errors. Furthermore, since the user does not have to memorize the subject names 9 and sentence numbers of the registered sentences, the burden on the user is greatly reduced.

When a user needs a sentence to refer to, he or she inputs a request sentence made up of synonyms, and the document processing device presents the most useful sentences one after another. Using a multi-window as shown in Fig. 7, the document being created af: is displayed on the left half of the CRT screen, and the subject name b1 of the text to be referenced and the retrieved candidate text b2 are displayed on the upper right part of the CRT screen. , the text C to be corrected selected from the searched candidate texts is displayed in the lower right portion of the CRT screen. In addition, the cursor exists in the lower right part, and you can insert characters at the cursor position,
Corrections are allowed. After the correction is completed, the text is moved to a predetermined position on the left half of the CRT screen.

On the other hand, when registering a text, as shown in Figure 8, use the multi-window as above and CR the created document a.
The subject name b1 of the registered text and the text b2 included in the subject name b1 are displayed on the upper right part of the CRT screen, and the text C that you wish to register from among the above documents a is displayed on the CRT screen. Display in the lower right part of the screen. This screen display allows the user to quickly register desired text with an appropriate subject name.

Effects> According to the present invention, a new sentence created by a user can be easily and accurately registered in the storage device by associating the keyword with the subject 1 sentence number, etc. This allows text to be retrieved quickly in response to a user's request.

[Brief explanation of drawings]

FIG. 1 is a block diagram of an embodiment of a document processing device according to the present invention, FIG. 2 is a block diagram of a conventional document processing device, FIG. 3 is a flowchart of an algorithm related to text retrieval, and FIG. FIG. 5 is an explanatory diagram of a kanji dictionary, FIG. 5 is an explanatory diagram of a sentence relationship table, FIG. 6 is a flowchart of an algorithm related to sentence registration, and FIGS. 7 and 8 are explanatory diagrams of a CRT screen. In the figure, 1: Character input section 2: Text search section 3: Document creation/correction section 4: Text storage section 6: Document output section
7: Sentence analysis unit 8: Dictionary 9: Sentence subject control unit 10: Sentence relationship table storage unit 11: Sentence improvement department agent Patent attorney Takeshi Sugiyama (and 1 other person) Figure 4 tσ (b)
(c) Figure 5 Figure 6 Figure 7 Figure 8 Procedural Amendments June 25, 1986

Claims (1)

[Claims] 1. Extracting means for extracting keywords symbolizing the meaning of words by analysis from words included in any text created by a user; a text subject control means for determining a text subject; a registration means for storing a text file related to the subject and registering the text by assigning a text number; 1. A document processing device, comprising: an index creation means for indexing a correspondence relationship between a user and a user, and is configured to provide an example of a sentence that matches a request sentence input by a user.