JPH0682363B2 - Document processor - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION <Industrial field of application> The present invention automatically classifies sentences created by a user or sentences entered from an electronic mail or the like, and creates a summary table showing an outline of the contents. The present invention relates to a document processing device that enables efficient document management by generating the document.

<Prior Art> Conventionally, in a Japanese word processor or a Western word processor which is a document processing device, when a relatively standard document such as a business document is created, a document created in the past can be used as it is or a part thereof is modified. It was often the case that a new document could be efficiently created simply by doing. It was also a big advantage of word processors.

Therefore, in order to easily find the desired sentence,
A name that sets the classification system in advance, classifies the documents created by the user based on the classification system, and symbolizes the contents of the document (document name)
Was stored on the auxiliary storage device.

<Problems to be Solved by the Invention> However, in order for a user to decide whether to classify a newly created document somewhere in the classification system, it was necessary to grasp all the classification systems. Moreover, when a plurality of users are using the same document processing device, all users must have the same criteria. Therefore, it is very difficult to classify documents uniformly according to their contents.

Also, regarding the documents classified into the same place, the number of similar documents increases as the number of stored documents increases, and the contents cannot be judged only by the document name. For example, documents with the document name “guidance” are roughly classified into a) party guidance, b) conference guidance, and c) large sale guidance, etc., and there are various documents with individual contents.

Therefore, in order to select what a user truly desires from many documents, it is necessary to check the document contents as well as the document name, which is very troublesome.

<Object of the Invention> The present invention solves the above-mentioned conventional problems, and when a document is stored in an auxiliary storage device, it is classified into concepts that symbolize the contents of the document and a summary table is automatically generated. It is therefore an object of the present invention to provide a new document processing device capable of efficiently performing the work of retrieving the document later.

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

FIG. 1 shows the basic configuration of the document processing apparatus of this embodiment.
The document processing device 1 displays the text input by the keyboard 2 on the display device 3, edits and creates the document, and stores it in the first auxiliary storage device 4 (for example, a floppy disk).

The keyword extracting means 6 extracts a plurality of words from the input sentence using the dictionary 5, and derives keyword candidates of a superordinate concept from the extracted word group.

The sentence concept control means 8 uses the keyword candidates extracted by the keyword extraction means 6 from the concept / keyword relationship table 7
Using, the keyword combination suitable for the concept candidate is determined from the concept candidate and the keyword candidate that symbolize the content of the sentence.

The keyword analysis means 10 analyzes the sentences before and after the word converted into the keyword in the sentence regarding a particularly important keyword described in the rule table 9 from the combination of keywords determined by the sentence concept control means 8. And extract the information.

The summary table generating means 12 uses the frame corresponding to the concept of the sentence selected by the sentence concept controlling means 8 from the frames stored in the frame accumulating section 11 to extract the sentence from the information extracted by the keyword analyzing means 10. If the result is successful, a summary table is generated and stored in the second candidate storage device 13 (for example, a floppy disk, which may be shared with the first auxiliary storage device 4). If unsuccessful, the same process is performed for the combination of the concept of the next candidate sentence and the keyword of the sentence concept control means 8.

The operation of this document processing apparatus will be described below.

For example, suppose that the user uses the keyboard 2 and inputs "Well, I sent the capacitor you ordered on January 2 by courier" ... (1). The document processing apparatus 1 divides the input sentence into word units, and gives the divided words to the keyword extracting means 6.
A conceptual diagram in which the dictionary 5 and the concept / keyword relationship table 7 are combined is shown in FIG. The keyword extracting means 6 uses the words "order", "condenser", and "feed" in the sentence to extract the dictionary 5 (10
Using 1), convert this word group into keyword candidates such as "order" → "order""request","capacitor" → "product", "send" → "dispatch""progress""ship". To do. Through the above processing, the keyword extraction means extracts six keyword candidates of “ordering, request, product, dispatch, progress, shipping”.

Next, the sentence concept control means 8 determines a combination of the concept candidate symbolizing the content of the sentence and the keyword related thereto from the extracted six keyword candidates by using the concept / keyword relationship table 7 (102). Here, the concept / keyword relationship table 102 is a series of keywords related to the concept of a sentence. For example, the concept “notice of adoption” is connected to the keywords “person, dispatch, request etc.”. With this concept / keyword relationship table, the concept candidate of the sentence is determined from the comprehensiveness of the keywords, and at the same time, the ambiguity in the dictionary 101 for polysemous words is eliminated. Through the above processing, the text concept control means 8 uses the “shipment notification” as a text concept candidate and the above 6 as a keyword combination.
From the one keyword candidate, three keywords “ordering, product, shipping” connected to “shipment notification” are extracted.

Next, in the keyword analysis means 10, regarding the keywords described in the rule table 9 among the extracted combinations of the above three keywords, the sentences before and after the word converted into the keyword in the sentence (1) are displayed. Analyze and extract information.

That is, in the rule table 9, for example, the rule such as IF THEN ([Order] + Given + X +) → (Position: Dealer, Ordered product: X) (2) is registered for the keyword “Order”. is there. This means that if a word in the sentence converted into the keyword "ordering" is followed by "Gift" and some characters followed by "", the position of the person who wrote this sentence is It is the position of the trader, that is, the side that received the order, and the character string X that is sandwiched between "Gived" and "O"
Means that there is a rule that indicates the contents of the ordered product, and this relationship is stored.

By the above processing, the keyword analysis means 10 extracts the context information such as order date → January 2nd position → supplier ... (3) ordered product → capacitor delivery means → courier service from the sentence (1).

Next, the summary table generation means 12 uses the text concept control means 8 to generate a summary table based on the information in (3) using the frame corresponding to the “notification of shipping” which is the concept candidate of the text (1).

For example, the “shipment notification” frame has the following structure.

"Notification of shipment" frame Position: Supplier Order date: Ordered item: (4) Delivery method: Delivery date: The left side of ":" in the frame shows the attribute name and the right side shows the value. Here, the constraint condition for the attribute name and the sentence having a value in advance to be the concept of the frame is shown. Also,
The attribute name that does not contain a value (in the example of (4), order date, ordered item, shipping means, shipping date) is the content that the text of the concept of the frame generally includes. These values are filled with the word information of the context information extracted by the keyword analysis means 10 to generate the summary table.

The generated summary table is stored in the second auxiliary storage device 13 after being classified into the concept of “shipment notification” class. Further, the address on the first auxiliary storage device 4 where the sentence (1) is stored is also stored at the same time.

When a user wants to retrieve a document, he or she specifies the desired concept, displays a plurality of summary tables included in the concept, and specifies the relevant contents from among them to make it suitable for the document requested from the address. Accurate and quick access.

<Effects of the Invention> As described above, according to the present invention, sentences created by the user or sentences input from e-mail or the like are automatically classified according to their contents, and a summary table showing an outline of the contents is attached. Since it is stored, the user can easily retrieve the desired document by checking the summary table, and efficient document management becomes possible.

[Brief description of drawings]

FIG. 1 is a block diagram showing an embodiment of the present invention, and FIG. 2 is a conceptual diagram in which a dictionary of the present invention and a conceptual keyword relation table are combined. In the figure, 1 ... Document processing device, 2 ... Keyboard, 3 ... Display device, 4 ... First auxiliary storage device, 5, 101 ... Dictionary, 6
…… Keyword extraction means, 7,102 …… Concept / keyword relationship table, 8 …… Sentence / concept control means, 9 …… Rule table, 10
...... Keyword analysis means, 11 …… Frame storage section, 12 ...
... summary table generating means, 13 ... second auxiliary storage device

Claims (1)

[Claims] 1. A keyword extracting means for extracting a keyword candidate symbolizing the meaning content of the word from a word included in an arbitrary sentence created by a user by analysis, and a concept candidate representing the content of the sentence and the keyword candidate. Sentence concept control means for determining a keyword combination suitable for the concept candidate from among, and before and after the word in the input sentence corresponding to an important keyword to express the meaning of the sentence in the keyword combination. Keyword analysis means for analyzing a sentence to extract context information of the input sentence, and a test as to whether or not the context information matches a frame composed of the attribute name of the concept candidate and a value corresponding to the attribute name. When a match is made, the concept of the sentence is determined from the concept candidates, and the value is filled with the word information of the context information to summarize the table. A summary table generation unit to be generated, a storage unit for storing an address of the summary table and the input sentence in association with the concept of the sentence, a document processing apparatus characterized by comprising a.