JP3845046B2 - Document management method and document management apparatus - Google Patents

Description

[0001]
BACKGROUND OF THE INVENTION
The present invention relates to a document management method and a document management apparatus. In particular, the present invention relates to a technique for associating a plurality of documents based on their contents.
[0002]
[Prior art]
In recent years, attention has been focused on technologies for centrally managing and effectively using a large number of files such as in-house documents created by different people. Accumulated documents are a condensation of knowledge and know-how, but in order to share and utilize such potential assets as truly meaningful information, managing documents in advance is a shortcut. The methods include a method of handling one document as one record from the beginning on the premise of creating a database, and a method of including a search index in each document.
[0003]
On the other hand, in order to efficiently extract a group of documents, there is a method of associating related documents in advance. Linking is realized by adding reference information of another document related to the document to an area such as a document header. For example, in the electronic mail software, when a reply mail is created for a specific received mail, the reference ID of the received mail is described in the header of the reply mail to maintain the relevance thereof. The reference relationship is also managed in the electronic bulletin board by a similar method.
[0004]
[Problems to be solved by the invention]
Since the document managed by each of the above methods is created in advance on the premise of a specific management method, processing such as addition of an index and reference information is performed at the time of creation. However, a document that has not been subjected to such processing is excluded from management by this method. In order to set these documents as management targets, it is necessary to manually perform an index adding process and a process of finding a reference relationship, which requires a lot of time and labor. On the other hand, even when documents are related to each other in a planned manner, the viewpoint of the given relevance is uniform, and it cannot always be used effectively effectively.
[0005]
The present inventor has made the present invention based on the above recognition, and an object thereof is to provide a technique for efficiently managing a large amount of document files. Another object of the present invention is to provide a technique for detecting a document type as a search key for a document file. Yet another object is to provide a technology that supports the construction of a document database. Yet another object is to provide a technology that supports analysis of trends in a plurality of documents. Yet another object is to provide a technique for visualizing the relationship between documents.
[0006]
[Means for Solving the Problems]
One embodiment of the present invention relates to a document management method. This method includes a step of determining a word / phrase as a viewpoint for finding relevance between documents, a step of extracting a plurality of documents including the determined word / phrase in content, and a step of associating the plurality of extracted documents with each other. Recognizing the presence of the context, and presenting a document with the recognized context.
[0007]
The “document” corresponds to a file generated by application software such as a word processor, presentation software, an e-mail client, and a scheduler. It may be text data generated by a server program such as CGI (Common Gateway Interface) of an electronic bulletin board. The “document” is assumed to be a data file mainly including text in order to be subjected to language analysis processing. However, the data format is not necessarily limited, and may be an image file, a CAD data file, or the like.
[0008]
“Phrase” includes words, phrases, phrases, etc., and the length of the phrase is not limited. A phrase composed of a plurality of words having a dependency relationship may be used. The “word / phrase” does not necessarily appear in the document, and may be a word / phrase representing a topic or concept common to the documents. The “context between documents” is a relationship between contents existing between documents, and may be a context of document contents when a plurality of documents are arranged in order of creation date / time or update date / time. The context is automatically detected based on the document content, and the feature is used as a search key for document search.
[0009]
The association of a plurality of documents may be a technique of forming a thread by setting a link with each other. A “thread” is a format in which a plurality of pieces of data distinguished for each topic are associated with data that precedes and follows in time series. In this aspect, a thread is formed for each word or phrase as a viewpoint. Generally, a thread format is used for an electronic bulletin board or an e-mail client, and a context between a series of documents in this aspect can be displayed in a thread format. “Time-series order” may be determined based not only on the date or time of creation or update of a document, but also on the date and time when the document was referred to or the date and time when the document passed the deliberation.
[0010]
According to this aspect, a plurality of documents to which no information is added regarding the relationship between documents are automatically associated. Since the detected relevance is based solely on the document content, the relevance chain indicates the context between the documents. Since this can be expressed in a familiar format such as a thread, it is possible to efficiently manage the entire document while grasping its relevance.
[0011]
Another aspect of the present invention relates to a document management apparatus. The apparatus includes a storage unit that stores documents, a condition setting unit that determines a word / phrase as a viewpoint for finding relevance between documents, and an extraction that extracts a plurality of documents including the determined word / phrase from the storage unit. A thread that recognizes the existence of a context between documents by associating a combination of adjacent documents with a link among a plurality of extracted documents and a sequence in accordance with a predetermined rule by a link and linking a plurality of links A management unit; and an output unit that presents a document in which the existence of the context is recognized.
[0012]
The storage unit mainly indicates a storage device such as a hard disk. The storage unit may be configured separately from the main body, and in that case, may be connected to the main body via a network. Various forms of storage positions and storage states of a plurality of documents can be assumed. The condition setting unit, the extraction unit, the thread management unit, and the output unit are realized by a central processing unit, a control device, and the like.
[0013]
The condition setting unit may determine a word / phrase as a viewpoint by automatically extracting a characteristic word / phrase included in the document content, or may be determined based on designation by a user. The “order according to a predetermined rule” may be, for example, ascending order of serial numbers assigned to each document, or an objective order may be derived by paying attention to some attribute included in each document. As the “link”, for example, a method such as XML Link in XML (eXtensible Markup Language) may be used to link the same phrases between documents.
[0014]
When this apparatus is realized as a file server, it is automatically associated with another document simply by transferring a new document to the server. Similarly, if past documents already accumulated in large quantities are input to the server, automatic association is processed. As a result, for example, files created or edited in large quantities by a plurality of people such as in-house documents can be efficiently managed without bothering the creator or the editor.
[0015]
Yet another embodiment of the present invention also relates to a document management apparatus. The apparatus includes a storage unit that stores documents, a condition setting unit that determines a word / phrase as a viewpoint for finding relevance between documents, and an extraction that extracts a plurality of documents including the determined word / phrase from the storage unit. And a thread management unit that recognizes the existence of a context between documents by linking a plurality of extracted documents and linking a combination of adjacent documents in chronological order by linking the links. And an output unit for presenting a document whose existence is recognized.
[0016]
Even in this mode, a large amount of files can be managed efficiently, and the causal relationship between documents can be presented in a form that can be grasped more intuitively by associating them in chronological order.
[0017]
This device includes both a case where it points to either a server or a user terminal installed on a network, and a case where it points to a system consisting of a server and a user terminal connected via a network. The functional blocks of the extraction unit, thread management unit, and output unit may be provided on the server side or the user terminal side. It may be provided in both the server and the user terminal, and each functional block in that case may have the same name. Each of these functions is provided in the form of a program module, and may be executed by either or both of the server and the user terminal, or may be downloaded from the server to the user terminal for execution.
[0018]
It should be noted that any combination of the above-described constituent elements, or the constituent elements and expressions of the present invention may be mutually replaced between methods, apparatuses, systems, computer programs, recording media storing computer programs, data structures, etc. This is effective as an embodiment of the present invention.
[0019]
DETAILED DESCRIPTION OF THE INVENTION
In the present embodiment, a plurality of documents are associated with each other based on the commonality of included words, and the context between documents is detected based on a plurality of relevance distributions and linkage modes. Furthermore, each document is classified according to the context mode, and this is handled as one of the search conditions. Thereby, relevance from various viewpoints can be found from a plurality of documents that are not associated in advance, and information serving as a substitute for the search index is generated from the tendency of the relevance. In addition, as long as words can be extracted from a document, they can be managed regardless of the data format.
[0020]
FIG. 1 is a diagram illustrating a configuration of a document management system according to an embodiment. The document management system 100 includes a plurality of user terminals 104 and a document management server 10 connected via a network 102. The document management server 10 may be configured to include a web server, and the user terminal 104 may be configured to include a personal computer and client software installed therein.
[0021]
The document management server 10 can be realized in hardware by elements such as a CPU of a computer, and is realized in software by a program having a language analysis function and a data management function. So, functional blocks that are realized by their cooperation are drawn. Therefore, these functional blocks can be realized in various forms by a combination of hardware and software.
[0022]
FIG. 2 is a functional block diagram showing the basic configuration of the document management apparatus according to the embodiment. The document management server 10 includes an input unit 12, a storage unit 14, a condition setting unit 16, a document management unit 22, an output unit 36, and a communication unit 42. Each unit transmits and receives data to and from a terminal on the network via the communication unit 42. The communication unit 42 may have a router function and a server function in addition to the data transmission / reception function.
[0023]
The storage unit 14 stores a plurality of documents. A document is stored in the storage unit 14 via the input unit 12. The input unit 12 is an interface for data input, and may process data input via another data port in addition to input via the communication unit 42. The input unit 12 may provide a user with a workspace such as a portal site or an electronic bulletin board for inputting and outputting documents. When the input unit 12 captures an external document, the user may specify a domain on the network as a capture range.
[0024]
The condition setting unit 16 has a function of determining a word / phrase (hereinafter referred to as a “view word / phrase”) as a viewpoint for finding relevance between documents, and includes a document analysis unit 18 and a word / phrase determination unit 20. The document analysis unit 18 analyzes the contents of the document stored in the storage unit 14 and extracts characteristic phrases. For example, a common phrase included in each document may be extracted, or a noun phrase including a proper noun may be extracted. A phrase may be extracted in a form including modifiers such as adjectives. Weighting may be performed according to whether or not the word is included in the document title, or the word may be extracted after the document is summarized. The priority of the extracted words may be set in advance.
[0025]
The document analysis unit 18 may determine a document range to be analyzed in accordance with the size of the workspace or division, or may analyze a document in a domain designated by the user.
[0026]
The phrase determination unit 20 determines the viewpoint phrase based on the extraction result from the document analysis unit 18 or the instruction from the user. The user instruction is received via the network 102 via the communication unit 42.
[0027]
The document management unit 22 includes an extraction unit 24, a time series processing unit 26, and a thread management unit 28. The time series processing unit 26 rearranges the plurality of documents stored in the storage unit 14 in time series. At this time, it may be rearranged based on time attributes such as creation date and update date and time of each document, or may be rearranged based on a history separately managed with respect to the history of document generation. The history may be recorded through a workspace provided by the input unit 12. The history may be stored in the storage unit 14.
[0028]
The extraction unit 24 extracts, from the storage unit 14, a plurality of documents that include the word / phrase determined by the condition setting unit 16. At this time, the extracting unit 24 combines documents adjacent to each other when linked in time series among a plurality of documents including the same word / phrase.
[0029]
The thread management unit 28 includes a link processing unit 30, a context recognition unit 32, and a classification processing unit 34. The link processing unit 30 associates viewpoint words / phrases that are common between the documents to be linked by describing them in the document by a link method such as XLink. The object to be linked may be based on an element surrounded by tags in the XML document. The context recognition unit 32 recognizes the existence of a context between a series of documents by linking a plurality of links as threads. Information for specifying a thread or context is sent to the classification processing unit 34.
[0030]
The classification processing unit 34 classifies a plurality of documents associated with a series of contexts according to the context mode recognized from each document. Each document usually has a plurality of contexts, and the context mode is classified into several types based on a certain tendency seen in the link form with the preceding and following documents. The type is saved as an attribute and handled so that it can be specified as one of the search conditions later. The attribute may be described in each document, or may be collectively saved in the storage unit 14 as management data. As a result, it is possible to narrow down the search and remove noise in an unprecedented form. The context type will be described later.
[0031]
The output unit 36 has a function of presenting a plurality of documents together with a context, and includes a search processing unit 38 and a display processing unit 40. The search processing unit 38 acquires a user designation regarding the context mode as one of the search conditions, extracts a document classified into the specified mode from the storage unit 14, and presents it as a search result. The display processing unit 40 has a function of sending data for displaying the recognized context on the screen in the form of a thread to the user terminal 104 and a function of sending a search result by the search processing unit 38 to the user terminal 104. The display processing unit 40 may provide a context correction function based on a user instruction through a screen for displaying the context.
[0032]
FIG. 3 is a diagram illustrating a context between documents linked by a phrase link included in each document. A plurality of contexts flow in the series of documents, and the documents are classified into several types according to the context modes recognized from the respective contexts. The classification processing unit 34 finds a feature point related to a context mode in each document. Specifically, the documents are classified based on whether the context mode corresponds to a start point, a merge point, a passage point, a branch point, or an end point in the document.
[0033]
In this figure, viewpoint words and phrases “Company A” and “Product B” are extracted from each document and links are set. Each document is arranged in time series. The words “Company A” and “Product B” are not extracted from the document before the first document 50 and are not linked to the previous document. Therefore, it is determined that the first document 50 is a type that has a strong tendency to become a context start point.
[0034]
The phrases “Company A” and “Product B” included in the first document 50 are linked to the phrases “Company A” and “Product B” included in the second document 52, respectively, which are further linked to the third document. 54 are linked to the words “Company A” and “Product B”. Therefore, it is determined that the second document 52 is a type that has a strong tendency to become a passage point of context.
[0035]
The phrase “Company A” included in the third document 54 is linked to the phrase “Company A” included in the fourth document 56, but the phrase “Product B” included in the third document 54 is the fifth The word “product B” included in the document 58 is linked. Therefore, the third document 54 is determined to be a type that has a strong tendency to become a context branch point. This type may be positioned as a document that has made a decision, or may be positioned as a possibility of being a minutes by further subdividing the branching mode.
[0036]
The phrase “Company A” included in the fourth document 56 and the phrase “Product B” included in the fifth document 58 are linked to the phrases “Company A” or “Product B” included in the sixth document 60, respectively. Is done. Therefore, it is determined that the sixth document 60 is a type that has a strong tendency to become a confluence of contexts. This type may be positioned as a document that is a convergence or milestone of the project.
[0037]
The phrases “Company A” and “Product B” included in the sixth document 60 are linked to the phrases “Company A” and “Product B” included in the seventh document 62, respectively, and are linked to subsequent documents. Not. Therefore, it is determined that the seventh document 62 is a type that has a strong tendency to become the end point of the context.
[0038]
Although only two of the first context 64 and the second context 66 are shown in the figure, in practice, many context recognitions are possible. There may be a case where the tendency of the aspect becomes clearer as the number of contexts to be recognized increases. The classification processing unit 34 in FIG. 2 may hold the classification criteria for such tendencies in advance.
[0039]
FIG. 4 is a diagram illustrating a state in which a plurality of documents arranged in time series are associated by a plurality of contexts. The vertical axis is the workspace axis, and the horizontal axis is the time axis. The workspace may be, for example, a space indicating the spread of each department or project in the company, or may be a space formed according to the classification of documents to be managed by this apparatus.
[0040]
The entire context space shown in the figure may be displayed on the screen of the user terminal 104. In that case, for example, when the user designates a specific viewpoint phrase, only the corresponding context may be highlighted. The user may select one of a plurality of contexts, and the display of the selected context may be switched to a tree-like thread form.
[0041]
FIG. 5 is a diagram illustrating an example of a screen that displays threads for each phrase in a tree shape. In the first thread 70, the title and creation date of each document included in the first context 64 of FIGS. 3 and 4 are displayed in a tree shape. Similarly, in the second thread 72, the title and creation date of each document included in the second context 66 are displayed in a tree shape. Thereby, a plurality of documents related to the phrase specified by the user can be visualized together with the chronological relationship.
[0042]
FIG. 6 is a flowchart showing processing executed by the document management apparatus according to the embodiment. A plurality of documents stored in the storage unit 14 are rearranged in chronological order (S10), and viewpoint words / phrases for finding a context between documents are determined (S12). A document to be linked including the viewpoint word / phrase is extracted (S14), and a link between words / phrases is added in a combination of documents that follow in time series (S16). A context in which a plurality of links are linked is recognized (S18), and the documents are classified according to the mode (S20). With the above processing, the preprocessing that is the premise of the search is completed.
[0043]
A search condition related to the classification of the document is acquired from the user together with the search keyword (S22), and the search is executed based on the search condition (S24). The document extracted as the search result is presented on the screen of the user terminal 104 (S26).
[0044]
The present invention has been described based on the embodiments. This embodiment is an exemplification, and it is understood by those skilled in the art that various modifications can be made to each component and combination of processing processes, and such modifications are within the scope of the present invention. Hereinafter, modifications will be described.
[0045]
In the embodiment, a link is set between documents containing the same word / phrase, but in a modified example, strictness is not required with respect to the identity of the word / phrase, the word / phrase is normalized for each document, a synonym dictionary or a controlled word dictionary May be used to provide flexibility.
[0046]
In FIG. 6, the process of rearranging the documents in chronological order is located before S12. However, as a modification, the process may be performed between S12 and S14. Alternatively, it is possible to adopt a procedure in which the rearrangement process is executed between S14 and S16, and the documents are rearranged in time series after extraction.
[0047]
In the embodiment, the context classification is handled as one of the search conditions, but in the modified example, the search is performed by adding a weight according to the number of contexts flowing in the document or a weight according to the relationship between the authority and the hub. You may prioritize the results.
[0048]
When posting data on an electronic bulletin board is input from the input unit 12, it is possible that a thread different from the thread managed on the electronic bulletin board is formed. On the electronic bulletin board, when a series of data that belongs to one thread is actually divided on the way, in the embodiment, it can be automatically divided into different threads.
[0049]
【The invention's effect】
According to the present invention, it is possible to improve the convenience of the user regarding the use of the accumulated information.
[Brief description of the drawings]
FIG. 1 is a diagram illustrating a configuration of a document management system according to an embodiment.
FIG. 2 is a functional block diagram showing a basic configuration of a document management server according to the embodiment.
FIG. 3 is a diagram showing a context between documents linked by a phrase link included in each document;
FIG. 4 is a diagram illustrating a state in which a plurality of documents arranged in time series are associated by a plurality of contexts.
FIG. 5 is a diagram illustrating an example of a screen that displays a thread for each phrase in a tree shape.
FIG. 6 is a flowchart showing processing executed by the document management server according to the embodiment.
[Explanation of symbols]
10 document management server, 14 storage unit, 16 condition setting unit, 20 phrase determination unit, 24 extraction unit, 28 thread management unit, 36 output unit.

Claims (3)

A step of determining a word / phrase as a viewpoint for the CPU to find relevance between documents;
CPU comprises the steps of extracting a plurality of documents containing phrase to the determined contents from the storage device, the CPU among a plurality of documents the extraction, together with associated by the time the link combinations of adjacent documents in sequential order, Recognizing the existence of a context between documents by chaining multiple links ;
An output means presenting a document in which the existence of the context is recognized;
Have a,
The recognizing step classifies the plurality of documents based on whether the context mode recognized from each of the plurality of documents corresponds to a start point, a merge point, a passage point, a branch point, or an end point in the document. Including
The presenting step includes a step of obtaining a user designation regarding the context mode, extracting a document classified into the specified mode from the storage device, and presenting the document. A storage unit for storing documents;
A condition setting unit that determines a word / phrase as a viewpoint for finding relevance between documents;
An extraction unit that extracts a plurality of documents including the determined word or phrase from the storage unit;
Among the plurality of extracted documents, a thread management unit that associates a combination of documents adjacent in chronological order with links, and recognizes the presence of a context between documents by linking a plurality of links;
An output unit for presenting a document in which the existence of the context is recognized;
Have a,
The thread management unit classifies the plurality of documents based on whether the aspect of the context recognized from each corresponds to a start point, a junction point, a passage point, a branch point, or an end point in the document,
The document management apparatus , wherein the output unit acquires a user designation regarding the context mode, extracts a document classified into the specified mode from the storage unit, and presents the document. Determining a word or phrase as a viewpoint for finding relevance between documents;
Extracting a plurality of documents containing the determined words in content;
Recognizing the existence of context between documents by associating a combination of adjacent documents in a chronological order among the extracted documents and linking a plurality of links ; and
Presenting a document in which the presence of the context is recognized;
To the computer ,
The recognizing step classifies the plurality of documents based on whether the context mode recognized from each of the plurality of documents corresponds to a start point, a merge point, a passage point, a branch point, or an end point in the document. Including
The presenting step includes a step of obtaining a user's designation regarding the context aspect, extracting and presenting a document classified into the designated aspect .

Images