JPH01304575A - Document processing device - Google Patents

Abstract

PURPOSE:To easily and efficiently draw up a document by accurately and efficiently retrieving a desired document after automatically classifying and registering generated documents on many sides. CONSTITUTION:A format control means 5 divides the document, which a user inputs, to plural sentences by paragraphs, and keyword candidates are extracted from words in these sentences by a keyword extracting means 7. A topic analyzing means 11 finds an important keyword expressing a topic of this document out of keyword candidates and performs retrieval for the validity in sentences. A document object classifying means 15 presents document object candidates based on keyword candidates and idiomatic expressions extracted from sentences. A document conception control means 18 checks matching between the topic and document object candidates to determine the combination between the topic and a document object as the document conception.

Description

DETAILED DESCRIPTION OF THE INVENTION <Industrial Application Field> The present invention relates to a document processing device that automatically performs format division and content classification of documents created by users or documents input from e-mails.

<Prior art> In document processing devices such as Japanese word processors,
When creating a relatively standard document such as a business document, it is often not possible to create a new document by using a previously created document as is or by making some modifications. There are 6 major advantages of Dobrocesoza.

Therefore, the user classifies the blade document that he or she has created according to a certain classification system, adds a classification index (document name) that symbolizes the sentence and IS content, and stores it in memory. A desired document is retrieved and recalled from memory using the classification index.

<Problems to be Solved by the Invention> However, the document classification method in the conventional word processor described above is based on manual operations based on the user's judgment. There is a problem in that it is necessary to have a complete understanding of the items that are not listed, and that each user has different criteria for determining items that are not on the list. Therefore, there is a drawback that it is extremely difficult to uniformly and efficiently classify created documents according to their contents. Furthermore, as the number of documents classified into the same category increases, the number of documents with similar names increases, making it impossible to clearly determine the document content based on the document name alone, making it impossible to conduct an accurate search. For example, documents with the document name "Letter of Congratulations" can be broadly categorized into (1) Letters of congratulations on the opening of a store, and (2) Letters of congratulations on the completion of a new building.
(3) There are letters of congratulations for promotions, promotions, etc., and the contents of each sentence are significantly different from each other.Furthermore, even though they are in the same category as above, for example, there are differences in the expressions of the preamble, main sentence, and concluding sentence for opening a store and opening a store. be.

Therefore, an object of the present invention is to have the document processing device itself automatically classify documents in a multifaceted manner based on the cause of occurrence, purpose of creation, format paragraph pattern, etc., and register them with a consistent classification index. An object of the present invention is to provide a new document processing device that can reliably and efficiently search for a desired document.

<Means for Solving the Problems> In order to achieve the above object, the document processing device of the present invention includes a format control means that identifies a format paragraph pattern and divides an arbitrary document input by a user into a plurality of sentences. , a keyword extraction means for analyzing the words contained in the above-mentioned sentences and extracting keyword candidates symbolizing the meaning of the words; and a keyword extraction means for expressing the topic that is the cause of the occurrence of the document from among the extracted keyword candidates. A topic analysis means that finds important keywords and analyzes the sentences preceding the words corresponding to the found keywords to check the validity of the topic, and extracts idiomatic expressions contained in the sentences.
A document purpose classification means that presents document creation purpose candidates based on the extracted idiomatic expressions and the keyword candidates, and consistency between the topics found by the topic analysis means and the document purpose found by the document purpose classification means. document concept control means for inspecting and determining a combination of topic and document purpose forming a document concept;

A document input by a user is divided into a plurality of sentences by first identifying the format paragraph pattern by the format control means, and the words in these sentences are analyzed by the keyword extraction means to extract their meanings. Keyword candidates that symbolize the are extracted. The Tobi-Tsuku analysis means selects, from among the keyword candidates extracted by the keyword extraction means,
Keywords important to express the topic that is the cause of the occurrence of the document are found, and the validity of the topic is examined by analyzing the sentences preceding the words corresponding to the found keywords. On the other hand, the document purpose classification means extracts idiomatic expressions included in the text, and presents document creation purpose candidates based on the extracted idiomatic expressions and the keyword candidates found by the topic analysis means. lastly,
The document concept control means inspects the consistency between the topic determined to be valid by the topic analysis means and the document creation purpose presented by the document purpose classification means, and determines a combination of topic and document purpose that forms a document concept. . The topic name 1 document purpose name determined in this way is added as a classification index to each of the divided sentences and stored in the storage device. At the time of inspection, sentences corresponding to the given topic name etc. are stored in the storage device. This allows new documents to be easily created.

<Example> Hereinafter, the present invention will be explained in detail with reference to illustrated embodiments.

FIG. 1 is a schematic diagram of the configuration of a Japanese word processor which is an example of the document processing device of the present invention. Editing is done by converting kanji etc.
While automatically classifying documents according to document content, input document data and processed document data are displayed on the display device 3 and stored in the auxiliary storage device 4.

The central processing unit 1 converts the input document data into paragraph structure, inden! - Format control means 5 that divides into preamble, main sentence, concluding sentence, etc. according to a format paragraph pattern such as an opening sentence, and selects a plurality of words from the divided main sentences while referring to a keyword dictionary 6, and selects the top words of the selected words. Keyword extracting means 7 extracts conceptual words, that is, keyword candidates, finds important keywords expressing the topic that is the cause of the occurrence of the document from the extracted keyword candidates, and extracts the sentence preceding the word corresponding to the found keyword. A topic analysis means 11 is provided for analyzing and checking the validity of the topic.

Further, the central processing unit 1 includes an idiomatic expression extraction means 1.
3 refers to the idiomatic expression dictionary 12 and determines document creation purpose candidates based on the idiomatic expressions extracted from the text and the extracted keyword candidates;
From among the l-picks that were determined to be appropriate by the topic analysis means II, a final selection is made with reference to the topic hierarchy relationship table 16 (see Figure 4), and the selected topics and the document purpose classification means 15 Document concept control means 18 is provided which inspects the consistency with the creation purpose candidate determined in step 1 with reference to the topic/document purpose correspondence table 17 (see FIG. 5) and finally determines a valid combination of the two.

In order to find important keywords in the topic analysis means 11, the topic classification means 9 first determines topics suitable for the keyword candidates extracted by the keyword extraction means 7 based on the topic/keyword relationship table 8;
This is done by selecting particularly important keywords listed in Rule Table 10 from among the determined combinations of keywords and topics. Further, the document purpose classification means 15 refers to the document purpose/keyword/idiomatic expression relationship table 14 based on the idiomatic expressions extracted by the idiomatic expression extraction means 13 and the keyword candidates extracted by the keyword extraction means 7. to determine the document creation purpose candidates. Then, the central processing unit l adds each topic name, document purpose name, and format classification name finally determined by the document concept control means 18 as an index to each sentence paragraph of the main text data divided by the format control means. Then, the information is stored and registered in the auxiliary storage device 4. On the other hand, the central processing unit l
searches the auxiliary storage device 11 for a document corresponding to the index inputted from the input device 2 at the time of search, and displays the document as the search result on the display device 3.

The automatic document classification process performed by the Japanese language word processor configured as described above will now be described.

When a user inputs a document from the input device 2, the central processing unit 1 divides the input document into 1 sentence, main sentence, end sentence, etc. as shown in FIG. , perform the following analysis to determine the document concept for the divided main sentences. Now, suppose that the main sentence that has been divided is the sentence 20 shown in Figure 3. Note that the vertical lines in the figure indicate references by the keyword term 11Y or the relational table. Then, the keyword extracting means 7 extracts the sentence 2 using the keyword dictionary 6.
Convert each word in 0 to a keyword. In this example,
As shown in the column of keyword 21 in the figure, [3 companies,
Seven keywords are extracted: transformation 1 store, position, appointment, and yIj. Next, the topic classification means uses the l-pick-keyword relationship table 8 from the seven extracted keywords to select "change" and r as topic candidates 22.
<f fJ Select J and select topic candidate [-change]
The keyword ``reform'' is extracted for the topic HLIR appointment, and the keywords ``post'' and ``appointment'' are extracted for the topic HLIR appointment.

Next, the l-pick analysis means II analyzes the extracted topic candidates? Il?・Rule table 1O of keyword combinations
For the keywords described in , analyze the sentence before the word corresponding to the keyword and write L. Pick Hou?
11i. In the above rule 510, for example, the following considerations regarding the keyword "change" are registered.

IF THEN([
Organization] + If there is a word such as )・big ``change'', it means that you should interpret ``change'' as a topic 1 personnel change, which is different from a simple change of address or phone number. Also, regarding the topic "Inauguration", rules have been registered, such as that the word converted to the keyword "Inauguration" must be preceded by the word (Director) converted to the keyword "Date position", and these rules Topic candidates are carefully selected based on this.

Next, the idiomatic expression extracting means 13 extracts idiomatic expressions included in the sentence that symbolize the purpose of the sentence using the idiomatic expression dictionary I2. In the example in Figure 3, as shown underlined in sentence 20, ``Unfortunately, I was told that I was given J and 1.''
is extracted as an idiomatic expression unique to letters. Then, the document purpose classification means 15 uses the document purpose/keyword and idiomatic expression relationship table 14 to determine that the above two idiomatic expressions are expressions that will eventually appear in the document purpose "congratulation letter" or "congratulation letter". Assuming that the keyword "Keij" indicates the sentence W purpose "congratulatory letter", the first candidate for the document purpose is determined to be "congratulatory letter" as shown in the column of document purpose candidates 23 in FIG.

Further, the document concept control means I3 selects a topic hierarchy relationship table I6 shown in FIG. 4 from among the carefully selected topic candidates.
, and shortlist the more specific topics. In the case of the example sentence, Topic 6, ``Inauguration'' is more specific (lower concept) than Topic 1, ``Personnel change'', so the topic is determined to be ``Inauguration''. Next, the validity of the combination of topic candidate and document purpose candidate is checked with reference to the topic/document purpose correspondence table 17 shown in FIG. In the example sentence, the combination of Topink ``Inauguration'' and document purpose ``Letter of Congratulations'' is considered appropriate, but in the case of the combination shown by the X in Figure 5,
Candidates for topics and document purposes will be reconsidered as inappropriate.

In this way, the topic name and document purpose name may be, for example, "Inauguration",
When it is finally determined that it is a “congratulatory letter,” the central processing unit 1 adds the topic name and the document purpose name as a classification index to each paragraph of the main text data divided by the format control means 5, and adds the above-mentioned topic name and document purpose name as a classification index. t is stored in the device 4 and registered. - When creating a new document using the document registered in this way, the user inputs the desired topic name and document purpose name from the input device 2 as inspection conditions.

Then, the central processing unit 1 stores the document data with the classification index corresponding to this search condition in the Nli auxiliary storage device 4.
, and display it on the display device 3.

Therefore, the user can easily and efficiently create a new document by modifying or referring to the displayed document.

In the above embodiment, each processing means 7.1 +, 15.18 provided in the central processing unit 1 is provided with various dictionaries 6.12 and tables 8.9° 14.16.1 for streamlining the processing.
7 and auxiliary processing means 9.13, the automatic classification of documents can be made even faster.

<Effects of the Invention> As is clear from the above description, the document processing device of the present invention includes a format control means for dividing a document input by a user into paragraphs into a plurality of sentences, and keyword candidates from words in this sentence. a keyword extraction means for extracting, a topic analysis means for finding important keywords expressing the topic of the document from the keyword candidates and searching for their validity in the text, and idiomatic expressions extracted from the keyword candidates and the text. document purpose classification means for presenting document purpose candidates based on the document purpose candidates; and document concept control means for determining the combination of topic and document purpose forming a document concept by checking consistency between the topic and the document purpose candidates. The determined topic name/document purpose name is added to the divided sentences as a classification index and stored, while the document can be created by searching for sentences that correspond to the given classification index. After automatically classifying and registering created documents quickly and multifacetedly, a desired document can be accurately and efficiently searched, and a new f2 document can be easily and efficiently created using this as a reference. It greatly contributes to dramatically increasing the efficiency of document management.

[Brief explanation of the drawing]

Fig. 1 is a schematic diagram of the configuration of a Japanese word browser that is an embodiment of the present invention, Fig. 2 is a diagram showing an example of format division according to the above embodiment, and Fig. 3 is a concrete example of the Mangoro processing of the above embodiment. Conceptual diagram, Figure 4 is a diagram showing an example of the hierarchical relationship of topics,
FIG. 5 is a diagram showing an example of whether a topic and a document purpose can be combined. I: Central processing unit, 2: Input device, 4: Auxiliary storage device, 5: Format control means, 7: Keyword extraction means, 11: Topic analysis means, 15: Document purpose classification means , +8--Document concept control means. Patent Applicant Sharp Aji Shiki Company Representative Patent Attorney Aoyama Haka16 Figure 2

Claims (1)

[Claims] (1) Format control means that divides any document input by the user into multiple sentences by identifying format paragraph patterns, and keywords that analyze words contained in the sentences and symbolize the meaning and content of the words. Keyword extraction means extracts candidates, finds important keywords to express the topic that is the cause of the document from the extracted keyword candidates, and analyzes the sentence preceding the word corresponding to the found keyword. and a document purpose classification means that extracts idiomatic expressions contained in a text and presents document creation purpose candidates based on the extracted idiomatic expressions and the keyword candidates. and document concept control means for checking consistency between the topic found by the topic analysis means and the document purpose found by the document purpose classification means, and determining a combination of topics and document purposes forming a document concept. , divides the input document, adds a topic name and document purpose name to each divided sentence as a classification index, and stores it in the storage device, while also extracting sentences from the storage device based on the given topic name, etc. A document processing device that can search and create documents.