JPH02297282A - Document processor - Google Patents

Abstract

PURPOSE:To efficiently perform a document processing by automatically extracting and registering a bibliographic item from an inputted document, also automatically extracting and registering the bibliographic item of a reference document described on the document, and performing the retrieval of the bibliographic item from those information. CONSTITUTION:A document selector 13 performs document selection by displaying a content stored in a bibliographic item memory device 4 on a document display device 14, and reads out document information selected by a user from a document memory device 3, and displays it on the device 14. At such a case, the selector 13 is connected in parallel with a reference relation traverse device 15 and an instruction selector 1. And the device 15 displays information with respect to the document or a document group to refer to the document displayed on the document display device 14, or the document or the document group to refer to the document being displayed at that time, and performs a control processing so as to display displayed document information on the device 14.

Description

[Detailed description of the invention] Industrial applications The present invention relates to a document processing device.

Conventional technology Conventionally, regarding document information retrieval, etc., for example, Japanese Patent Application Laid-open No. 63-
There is a patent information search device as shown in Japanese Patent No. 123164. In this method, patents related to a certain patent are stored in a related patent number storage means using a patent number and a keyword as clues, and this is used to search for related patents.

Problems to be Solved by the Invention However, such a method can only be applied to a limited amount of information such as patent information. That is, general papers and the like do not have information such as patent numbers that enable unique identification, and must be searched based on bibliographic information such as work names and titles. In particular, for research papers, it is necessary to search for papers, etc. that the paper refers to in a so-called "Imozuru-style" manner. Therefore, there is a need for an apparatus that is versatile enough to handle such cases and that can efficiently perform document searches.

Means for Solving the Problem A document input means, a document storage means, a document display means, a bibliographic item storage means, a reference document list storage means, and a method for extracting bibliographic items such as author name and title from an input document. bibliographic item registration means for registering in bibliographic item storage means; reference document list registration means for extracting bibliographic items such as author names and titles of reference documents from input documents and registering them in said reference document list storage means; and said reference document list. and a reference document search means for searching a reference document using the bibliographic item storage means based on registered information in the storage means.

When a document is input by the action document input means, first, bibliographic information such as the author's name of the document is automatically extracted by the bibliographic registration means and registered in the bibliographic storage means. At the same time, information on bibliographic items such as titles is automatically extracted from the list of reference documents such as references and cited documents that are displayed at the end of the input document, such as the end of the paper. Register it in the list storage means.

Therefore, the next step is to use the reference document search means to search for the bibliographic information of the document that has already been input and registered in the bibliographic information storage means based on the registered information in the reference document list storage means, that is, author name, title, etc. As a result, the reference document is searched from among the registered documents. At this time, since the search is performed using bibliographic items such as author names and titles, it is possible to perform a versatile document search without the need for management numbers such as document numbers that correspond to patent numbers.

Example An embodiment of the present invention will be described based on the drawings.

First, a document input means 2 is provided which is activated by a user's instruction selection using an instruction selection means 1 when registering a document. The document input means 2 accepts text from a keyboard or the like, output text from an OCR device, or text via communication, and stores the information of the input document in the document storage means 3. The document input means 2 also performs document understanding of the input document, automatically extracts bibliographic information such as author name, title, etc., and registers information on the extraction results in the bibliographic information storage means 4. A bibliographic item registration means 4 is connected. The document understanding etc. in this bibliographic item registration means 4 can be performed, for example, in the above-mentioned Japanese Patent Application Laid-Open No.
This may be carried out in accordance with the document understanding operation shown in Publication No. 64. Furthermore, a reference document list registration means 6 is provided, into which information on documents input by the document input means 2 is inputted via the bibliographic item registration means 5. This reference document list registration means 6 searches for a character string such as ``Para reference bibliography'' or ``Para cited bibliography'' that exists at the end of the input document, detects a bibliography list of reference documents that follows, and Bibliographic information such as the author name and title of a document is registered in the reference document list storage means 7.

A reference document search means 7 is connected to the reference document list registration means 6. The process itself in the reference document retrieval means 8 conforms to the document understanding method disclosed in the above-mentioned publication. The reference document search means 8 reads pairs of author names and titles one by one from the reference document list that has just been registered in the reference document list storage means 7, and matches them with the pairs of documents that have already been registered. In order to check whether the bibliographic information exists in the bibliographic item storage means 4, the bibliographic information is compared with the record registered in the bibliographic item storage means 4. If the verification is successful, a link indicating a reference/referenced relationship is set between the documents by the reference relationship link setting means 9.

Specifically, this can be realized by using a pointer mechanism as illustrated in FIG. 2, which will be described later. If the matching is not successful, it means that the document with the pair of author name and title has not been input into this device yet, so information about that document is registered in the unregistered document information register. The unregistered document information is registered in the unregistered document information storage means 11 using the means 10 (.

Next, the matching process is performed for other author name and title pairs in the reference document list in the same way, and when the process for all author name and title pairs is completed, this time, the reference source document search means Control is then transferred to the search process by No. 12. This reference source document retrieval means 12 performs processing on the input current document in order to cope with the case where the referenced document is already stored. Specifically, the records in the already registered unregistered document information storage means 11 are read out one by one, and it is checked whether the records match the author name and title of the current document. If the matching is successful in this case as well, the reference relationship link setting means 9 sets a link indicating a reference/referenced relationship between the documents, as in the case of the reference document search means 8. If it does not match, move on to matching with the next record. When all the records in the unregistered document information storage means 11 are checked, all processing operations at the time of document registration are completed.

Next, the operation when searching for a document will be explained. The search operation is performed by activating the document selection means 13 by a user's instruction selection using the instruction selection means 1. This document selection means 13 causes the document display means 14 to display the content stored in the bibliographic item storage means 4 for document selection, reads document information selected by the user from the document storage means 3, and displays the document information on the document display means 14. 14 is displayed. On this occasion,
In this embodiment, the document selection means 13 receives a user's instruction so that the control can be transferred to the reference relation traversing means 15 which searches the document based on the reference relation link.
The reference relationship traversal means 15 and the document selection means 13 are connected in parallel to the instruction selection means 1. This reference relationship traversing means 15 retrieves information about a document or a group of documents that refer to the document displayed on the document display means 14 at that time, or a document or a group of documents that the document displayed at that time refers to. When the user makes a selection regarding the displayed document information (bibliographic items), control processing is performed to display the selected document information on the document display means 14. As a technique for realizing this, there is an existing multi-window method.

In this way, according to this embodiment, efficient document search is possible. In particular, if the reference relationship link setting means 9 and the reference relationship traversal means 15 are used, reference destinations and referenced documents can be automatically found, making document retrieval easy and efficient.

Furthermore, by using the unregistered document information registration means and the unregistered document information storage means 11, unregistered documents can also be managed.
This also makes it easier to handle when adding documents.

FIG. 2 schematically shows an example of the document storage means 3, bibliographic item storage means 4, reference document list storage means 6, unregistered document information storage means 11, and reference relationship links.

Here, each storage means is actually realized using a computer memory or an external storage, and they are connected to each other by a pointer mechanism. In each process, consistent processing is possible by following these pointers. The memory includes a directory area 16, a text storage area 17 (corresponding to the document storage means 3), a bibliographic item storage area 18 (corresponding to the bibliographic item storage means 4), and a reference document list storage area 19 (reference document list storage area 19). (corresponding to means 6), referenced table area 20, and unregistered document list storage area 21 (corresponding to unregistered document storage means 11)
It is divided into The directory area 16 has a pointer field Pa to the text, a pointer field pb to bibliographic items, and a pointer field PC to the reference document list.

The text storage area 17 stores a plurality of document information.
The end of each document is distinguished by an EOF (end of file). The bibliographic information storage area 18 has a pointer field Pd to the directory and a pointer field Pe to the referenced table.
It also has an author name field A, a title field B, and an other field C. The reference document list storage area 19 has a pointer field Pf to a directory, an author name field D, a title field E, and a pointer field Pg to a reference destination.

The referenced table area 20 serves as a pointer to the reference source.

Now, to explain using the specific example shown in FIG.
・Various pointers related to the document with the title ``are stored. That is, the text information is located in the second sentence area in the text storage area 17, and the bibliographic information is located in the third area in the bibliographic storage area 18. Further, a reference document list for this document is in the reference document list storage area 19, and in the illustrated example, three documents are restored.

However, the first document titled "Mumu High School Baseball Club History J" written by Omutabe is an unregistered document and has not yet been stored in this device, so it is not stored in the unregistered document list storage area. 21, the bibliographic information is copied and stored as is. It should be noted that each unregistered document holds a link L in the form of a pointer mechanism to indicate what is referring to it. Here, it can be seen that the document ``called Maro◆・'' in the bibliographic storage area 18 is being referred to. Next, the second document in the reference document list storage area 19 titled "Study of High School Baseball" written by Omutabe has already been registered, as can be seen by checking the bibliographic information storage area 18. Link L1 using pointer
are tied together. Therefore, it is also possible to search for documents that refer to "research on high school baseball." That is, the bibliographic item storage area 18 has a pointer column Pe to the referenced table area 20, and by following this pointer, the pointer to the reference source can be found. In this way, bidirectionality of reference destination/reference source of various documents is realized. Although not particularly illustrated, reference relationship traversal can also be realized by using these pointers.

According to this embodiment, management of research papers in particular becomes easy.

This is because the reference/referenced relationships between papers are automatically expressed as links as papers are registered, so you can learn about trends in the field by looking at the links, and
This is because it becomes possible to perform a quick search.

Effects of the Invention The present invention automatically extracts and registers bibliographic items from an input document as described above, and at the same time automatically extracts bibliographic items of reference documents described in the document.
9. Based on this information, the system was configured so that related documents can be searched by searching the bibliographic information of documents that have already been entered, so it is possible to search using bibliographic information such as author name and title. This eliminates the need for management numbers such as document numbers that correspond to patent numbers, and enables versatile document searches.

[Brief explanation of drawings]

The drawings show an embodiment of the present invention; FIG. 1 is a block diagram, and FIG. 2 is a schematic diagram showing an example of stored contents. 2... Document input means, 3,17... Document storage means,
4.18... Bibliographic item storage means, 5... Bibliographic item registration means, 6... Reference document list registration means, 7, 19.
. . . Reference document list storage means, 8 . . Reference document search means, 14 . . . Document display means.

Claims (1)

[Claims] A document input means, a document storage means, a document display means, a bibliographic item storage means, a reference document list storage means, and bibliographic items such as author names and titles are extracted from the input document and registered in the bibliographic item storage means. bibliographic item registration means; reference document list registration means for extracting bibliographic items such as author names and titles of reference documents from input documents and registering them in the reference document list storage means; and registration information in the reference document list storage means. 1. A document processing device comprising a reference document search means for searching a reference document using the bibliographic information storage means based on the bibliographic item storage means.