JP2000276480A - Document processing method and apparatus, and recording medium - Google Patents

Abstract

(57) [Summary] [Problem] Find out past classification items and documents by going back date and time. SOLUTION: A classification model for classifying an electronic document, which is provided with information on an internal structure composed of a plurality of elements, into a plurality of classification items, and refers to an update date and time of the classification model recorded in the classification model. An input unit 20 for processing a document and inputting a date and time;
And a control unit 11 for searching for an electronic document based on the classification model corresponding to the date and time input in step (1).

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a document processing method and apparatus for processing an electronic document having an internal structure assigned to an element, and a recording medium on which a program for processing the electronic document is recorded.

[0002]

2. Description of the Related Art Conventionally, WWW (World Wide Web) has been provided as an application service for providing hypertext information in a window format on the Internet.

[0003] The WWW is a system for executing document processing for creating, publishing, or sharing a document, and showing the way of a new style document. However, from the viewpoint of practical use of documents, advanced document processing beyond WWW, such as classification and summarization of documents based on the contents of the documents, is required. For such advanced document processing, mechanical processing of the contents of the document is indispensable.

However, mechanical processing of the contents of a document is still difficult for the following reasons. First
HTML, a language that describes hypertext
(Hyper Text Markup Language) stipulates the expression of a document, but hardly specifies the content of the document. Second, a network of hypertexts formed between documents is not always easy for a reader of the document to understand the contents of the document. Third, although the author of a document generally works without the convenience of the reader in mind, the convenience of the reader of the document is not coordinated with that of the author.

[0005] As described above, the WWW is a system showing the way of a new document. However, since the document is not mechanically processed, advanced document processing cannot be performed. In other words, to perform advanced document processing,
Documents need to be processed mechanically.

Therefore, a system for supporting mechanical processing of documents has been developed based on the results of natural language research, with the goal of mechanical processing of documents. As a document processing based on natural language research, mechanical document processing using tags attached to a document has been proposed on the assumption that attribute information about the internal structure of the document by the author of the document, so-called tags are added. I have.

[0007]

By the way, with the recent spread of computers and the progress of networking, there has been a demand for sophisticated document processing. For example, as documents received via a network are stored in a computer, the amount of stored documents becomes enormous.

SUMMARY OF THE INVENTION The present invention has been made in view of the above-described circumstances, and is directed to a document processing method and apparatus for easily searching for a desired document by going back to the date and time from the stored document and from the stored document. It is an object of the present invention to provide a recording medium in which a document processing program for easily retrieving a desired document by going back date and time is recorded.

[0009]

In order to solve the above-mentioned problems, a document processing method according to the present invention provides a document processing method for processing an electronic document by referring to the update date of the classification model recorded in the classification model. The method includes an input step of inputting a date and time, and a search step of searching an electronic document based on a classification model corresponding to the date and time input in the input step.

A document processing apparatus according to the present invention is a document processing apparatus for processing an electronic document by referring to an update date and time of the classification model recorded in the classification model, wherein the input means inputs a date and time; Search means for searching for an electronic document based on the classification model corresponding to the date and time input by the means.

[0011] A recording medium according to the present invention is a recording medium in which a document processing program for processing an electronic document with reference to the update date and time of the classification model recorded in the classification model is recorded. , An input process for inputting a date and time, and a search process for searching an electronic document based on a classification model corresponding to the date and time input in the input process.

[0012]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Hereinafter, embodiments of a document processing method and apparatus and a recording medium according to the present invention will be described with reference to the drawings.

As shown in FIG. 1, a document processing apparatus according to an embodiment of the present invention includes a main unit 10 having a control unit 11 and an interface 12, an input unit 20 which receives an input from a user and sends it to the main unit 10. A receiving unit 21 that receives an external signal and sends it to the main unit 10, a display unit 30 that displays an output from the main unit 10, and a recording / reproducing unit 31 that records / reproduces information on / from a recording medium 32. Have.

The main body 10 has a controller 11 and an interface 12, and constitutes a main part of the document processing apparatus. The control unit 11 includes a CPU 13 that executes processing in the document processing apparatus, and an R that is a volatile memory.
It has an AM 14 and a ROM 15 which is a nonvolatile memory. The CPU 13 temporarily stores data, if necessary, in the RAM 14 according to a procedure recorded in the ROM 15, for example, and performs control for executing the program. The interface 12 includes an input unit 2
0, the receiving unit 21, the display unit 30, and the recording / reproducing unit 31 are connected. Under the control of the control unit 11, the interface 12 inputs data from the input unit 20 and the reception unit 21, transmits data to the display unit 30, and records / reproduces the data.
For data transmission / reception to / from, the timing of data transmission is adjusted or the format of data is converted.

The input section 20 is a section for receiving an input from the user to the document processing apparatus. This input unit 20
For example, it is composed of a keyboard and a mouse. The user can use the input unit 20 to input a keyword using a keyboard or select and input an element of a document displayed on the display unit 30 using a mouse. Here, the element is an element constituting a document, and includes, for example, a document, a sentence, and a word.

The receiving section 21 is a section for receiving a signal transmitted from the outside to the document processing apparatus via, for example, a communication line. The receiving unit 21 receives a plurality of documents that are electronic documents. The receiving unit 21 sends the received data to the main unit 10.

The display unit 30 displays the output from the document processing device. The display unit 30 is, for example, a cathode ray tube (CRT) or a liquid crystal display (liquid crytal).
stal display (LCD), which displays one or more windows and displays text,
For example, displaying figures.

The recording / reproducing section 31 records / reproduces information on / from a recording medium 32 such as a so-called floppy disk. In the recording medium 32, a document processing program for processing a document is recorded. This recording medium 32 will be further described later.

Next, a document according to this embodiment will be described. In the present embodiment, a document has its internal structure described by attribute information by tagging. Document processing in the document processing device is performed with reference to a tag attached to the document. In the present embodiment, a semantic / pragmatic tag that enables mechanical understanding of the content of a document between multiple languages is added together with a syntactic tag indicating the structure of the document. Has given to the document.

In this embodiment, syntactic tagging describes the internal structure of a document.
The internal structure by tagging is as shown in FIG.
Each element such as sentence, vocabulary element etc. is usually a link,
It is configured to be linked by a reference / referenced link. In the figure, a white circle “○” indicates an element, and the lowest white circle is a vocabulary element corresponding to the minimum level word in the document. A solid line indicates a normal link (normallink) indicating a connection between elements such as a document, a sentence, and a vocabulary.
). A broken line is a reference link indicating a dependency relationship between a reference and a referenced. The internal structure of a document is document, document,
Subdivision, paragraph (paragrap)
h), sentence (sentence), subsentential segment (subsentential segment),..., vocabulary elements. Of these, subdivisions and paragraphs are optional.

In the present embodiment, as the semantic / pragmatic tagging, there is a method of describing information such as a meaning like a meaning of a polysemous word. Tagging in the present embodiment is based on an XML (Extended Markup Language) format similar to HTML (Hyper Text Markup Language).

An example of tagging is shown below, but tagging a document is not limited to this method. In the following, examples of English and Japanese documents are shown, but the description of the internal structure by tagging can be similarly applied to other languages.

For example, "Time flies like an arro
The following tag can be attached to the sentence "w."

<Sentence><noun phrase word meaning = “time0”> time </ noun phrase><verbphrase><verb word meaning = “fly1”> flies <// verb><adjective verb phrase><adjective verb word meaning = like0> like 〈adjective verb〉 〈noun phrase〉 an 〈noun word sense ＝ “arrow0”> arrow 〈/ noun〉 〈／ noun phrase〉 〈／ adjective verb phrase〉 〈／ verb phrase〉. , <Noun>, <noun phrase>, <verb>, <verb phrase>, <adjective verb>, <adjective verb phrase> are prepositional phrases including sentences, nouns, noun phrases, verbs, verb phrases, adjectives, respectively. It represents a syntactic structure of a sentence such as a postposition phrase / adjective phrase or an adjective phrase / adjective verb phrase. Tags are arranged corresponding to immediately before the head of the element and immediately after the end of the element. Here, the tag placed immediately after the end of the element indicates that it is the end of the element by the symbol “/”. This element indicates syntactic constructs, ie, phrases, clauses, and sentences. In addition,
The meaning = “time0” is a plural meaning of the word “time”,
That is, it indicates the 0th meaning of the plural meanings. Specifically, “time” has a noun and a verb, but here, the word “time” indicates that it is a noun. Similarly, the word "orange" has the meaning of color and fruit, which can also be distinguished by meaning.

The document according to the present embodiment can display a syntactic structure in a window 101 of the display unit 30 as shown in FIG. In this window 101, vocabulary elements are displayed on the right half 103, and sentence structures are displayed on the left half 102, respectively.

In this window 101, the following document "Mr. A's B"
In C City, where the event ended, some popular and general newspapers have announced on their papers that they will voluntarily regulate their photographic coverage. Is displayed. An example of tagging this document follows.

<Document><sentence><adjective verb phrase relation = “position”><nounphrase><adjective verb phrase place = “C city”><adjective verb phrase relation = “subject”><noun phrase identifier =
“B meeting”><adjective verb phrase relation = “affiliation”><person name identifier = “Mr. A”> Mr.
<Person name></ adjective verb phrase><organization name identifier = "B meeting"
> B meeting </ organization name></ noun phrase></ adjective verb phrase> finished </ adjective verb phrase><place name identifier = "C city">
C city </ place name></ noun phrase>, </ adjective verb phrase><adjective verb phrase relation = "subject"><noun phrase identifier = "pres
s ”syntactic =“ parallel ”><nounphrase><adjective verb phrase> Some </ adjective verb phrases> popular paper </ noun phrase> and <noun> general paper </ noun></ noun phrase> / Adjective verb phrase><adjective verb phrase relation = “object”><adjective verb phrase relation = “content” subject = “press”><adjective verb phrase relation =
"Object"><nounphrase><adjective verb phrase><noun co-reference =
"B-kai"> so </ noun></ adjective verb phrase> photo coverage <
</ Noun phrase></ adjective verb phrase> Self-regulating </ adjective verb phrase> Policy </ adjective verb phrase><adjective verb phrase relation = “position”></ adjective verb phrase> on paper. </ Sentence></document> In this document, "part of popular paper and general paper" is expressed in parallel by a tag of syntactic = "parallel". The definition of parallel is to share a dependency relationship. If nothing is specified, for example, <
Noun phrase relation = x><noun> A </ noun><noun> B <
/ Noun></ noun phrase> indicates that A has a dependency on B. Relation = x represents a relation attribute.

Relational continuity describes the interrelationships between syntactic, semantic, and rhetorical. Grammar functions such as subjects, objects, and indirect objects, subject roles such as an actor, a subject, a beneficiary, etc., and rhetorical relations such as a reason and a result are described by the relation attributes. In the present embodiment, the subject,
Describe relational attributes for relatively easy grammar functions such as object and indirect objects.

In this document, "Mr. A"
For proper nouns such as "B Association" and "C City", place names,
Attributes are described by tags such as person names and organization names. These words to which tags such as place names, personal names, and organization names are given are always proper nouns.

The receiving section 21 of the document processing apparatus receives a document having an internal structure by tagging. The receiving unit 21 sends the received document to the main body 10. Documents sent to the main unit
For example, it is stored in the RAM 14. As described above, the received document is stored in the document processing apparatus. The user can cause the display unit 30 to display a desired document by inputting to the input unit 20.

On display unit 30, the document is displayed on a window whose size can be changed, for example. The document can be displayed by arranging a plurality of documents side by side by a plurality of windows, or by overlapping a plurality of windows. Also, the display of the document summary can be replaced with the display of the document.

The control unit 11 of the document processing apparatus performs various processes on the document displayed on the display unit 30 according to the input to the input unit 20 by the user. The input to the input unit 20 by the user is performed by specifying the area of the display unit 30 by clicking the mouse of the input unit 20 or by inputting a keyword using the keyboard of the input unit 20. The display unit 30 displays a cursor linked to the mouse of the input unit 20.

The document processing apparatus creates feature information representing the features of each document, that is, an index, from the plurality of received documents. With reference to FIG. 4, creation of an index will be described.

In step S11, the receiving section 21 of the document processing apparatus receives a plurality of externally transmitted documents.
The document processing device stores the plurality of documents received by the receiving unit 21 in, for example, the RAM 14 under the control of the control unit 11.

In step S12, the control unit 11 of the document processing apparatus creates an index by extracting the characteristics of each document. That is, for example, the control unit 11
Then, a plurality of documents received in step S11 stored in step S11 are read and an index of each document is created. The control unit 11 stores the created index in the RAM 14
To memorize.

The index represents the characteristics of each document. The document processing apparatus automatically classifies documents by referring to the index of each document based on the classification model of the document classification. The classification model includes a plurality of classification items, that is, categories, for classifying documents. Each category has a category index representing its characteristics.

In step S13, the user browses the document displayed on the display unit 30 of the document processing device.
That is, the control unit 11 of the document processing apparatus controls the display unit 30 to display a desired document from a plurality of stored documents as desired by the user. If the entire document cannot be displayed due to the size of the display area of the window, a part of the document is displayed in the window. The selection of a desired document is performed by the user selecting with the input unit 20. Step S13 is provided as needed by the user. Also, in the figure, this step S14
Is represented by a parallelogram, which corresponds to a user operation. The same applies to the following.

Here, a specific example of display on the display unit 30 will be described. This example is a graphic user interface (GU) that allows the user to freely set and change categories for classifying documents.
I). In this GUI, documents are automatically classified according to the category set by the user.

As shown in FIG. 5, the GUI window 301 includes an operation button 302, a first category display section 303 for displaying "sports news", and a second category display section for displaying "business news". Each category is displayed, such as a third classification display section 305 for displaying “political news” 304. The classification display section corresponding to each category displays the title of the document and the first part of the document.

In this window 301,
The operation button 302 is used to reset the state of the window on the screen to an initial position (position reset).
And a button for a browser that invokes a browser for reading the contents of the document, and a button for exiting from this window. Note that the size of each of the above-described classification display sections is not fixed but can be changed. The title or label of each category display can also be changed.

The automatic classification of documents aims at responding to the interests of the user and improving the efficiency when the user searches for a document by determining the category according to the individual request of the user.

In this window 301, in addition to the classification of the document at the present time, it is possible to access the category of the classification that the user has performed in the past and the documents belonging to that category by going back to the past date and time. Here, “date and time” refers to both date and time. The same applies to the following.

The control unit 11 of the document processing device includes an input unit 20
In accordance with the input of the past date and time from, the updated classification date and time recorded in the classification model are referred to to search for a matching classification model. Then, the control unit 11 reproduces the classification status of the retrieved classification model in the GUI window 301 of the display unit 30.

For example, the window 301 of this GUI
According to the classification model of the past date and time desired by the user, “sports news” in the first classification display unit 303,
“Business News” in the second classification display unit 304, third
The document is displayed in each category such as “political news” of the category display unit 305 based on the past category model.

Further, the user can search for a document by going back in time for each category such as "sports news" in the first classification display unit 303, for example. this is,
This is performed by referring to the update date and time of each document recorded in the category index. Thus, the document processing device
By going back in time, it provides the convenience of accessing the category of the past classification and the documents belonging to the category.

In step S14, the user creates a category for a plurality of documents browsed on the display unit 30 of the document processing apparatus in step S13, and classifies the plurality of documents. In the document processing apparatus, a category for classifying documents is set, for example, by adding, changing, or deleting a desired category in a window having an area divided according to the number of categories. . The classification of a plurality of documents into categories is performed by, for example, dragging an icon of a document displayed in a window to move the document to a desired area. The newly created category and the result of the operation are stored in the RAM 14, for example. For more information on creating categories and classifying documents,
Further details will be described later.

In step S15, the control unit 11 of the document processing apparatus executes the creation of the category performed in step S14 and the creation of a classification model based on the classification operation according to this category. The document processing apparatus reads, for example, the result of the category and classification operation in step S14 stored in the RAM 14. Then, based on the result, the control unit 11 of the document processing apparatus performs, for the plurality of documents classified into each category,
Characteristic proper nouns in each category, meanings other than proper nouns,
The document addresses of the classified documents are collected to generate a classification model.

The reason why the meaning is used instead of the word itself in cases other than proper nouns is that the same word may have a plurality of meanings. Then, the control unit 11 of the document processing apparatus stores the classification model thus created in, for example, the RAM 14. The details of the generation of the classification model will be described later.

In step S16, the control unit 11 of the document processing device registers the classification model created in step S15. Then, this series of steps ends.

For example, for the classification item created in step S14 in FIG. 4, a process of changing the classification model manually by the user is performed as shown in FIG. In the manual classification of documents, the user manually updates the classification destination of the category for classifying the document received in step S11, for example, and the category itself in the first step S17.
This manual update is performed by the user operating a GUI window 301 shown in FIG.

In step S17, the user changes the category for classifying a plurality of documents in the window 301. For example, the user may select window 301
The category itself can be changed by newly installing or deleting the classification display section displayed in the section. Further, the user can change the category to which the document is classified by moving the document displayed on each classification display unit of the window 301 to another classification display unit.

In step S18, the control unit 11 determines, based on the classification destination of the document into the category and the change in the category itself, input by the user in step S17.
Perform processing to change the classification model. Then, in step S19, the control unit 11 registers the changed classification model. For example, the control unit 11 causes the RAM 14 to store the changed classification model. The update date and time of the classification model is also registered at the same time. The classification model before updating is also stored.

Referring to FIGS. 4 and 6, a classification model serving as a reference for classifying documents is created and updated.
Automatic classification of documents can be performed based on this classification model.

Referring to FIG. 7, the operation of automatic document classification performed by the document processing apparatus will be described. Step S2
In 1, the receiving unit 21 of the document processing apparatus receives a new document transmitted from the outside via, for example, a communication line.
Since the operation of receiving a document in the document processing apparatus has been described in detail in step S11, the description is omitted here. The received document is stored in the RAM 14, for example.

In step S22, the control section 11 of the document processing apparatus reads the document received in step S21 stored in the RAM 14, for example. The control unit 11
An index is created by extracting words representing the characteristics of each document from the new document. The index records the reception time of the document. And the control unit 1
1 stores the index in the RAM 14, for example.

In step S23, the control unit 11 of the document processing apparatus classifies each indexed document into one of a plurality of categories based on the classification model.
Then, the control unit 11 stores the classification result in, for example, the RAM 1
4 is stored. The details of such automatic classification of documents will be described later.

In step S24, the control section 11 of the document processing apparatus reads the result of the classification in step S23 from, for example, the RAM 14, and updates the classification model based on the result of the automatic classification.

In step S25, the control section 11 of the document processing apparatus registers the result of the classification model update in step S24. At this time, the update date and time of the category are stored, and the classification model before the update is stored without being deleted.

The document model indicates that a document is newly distributed,
When a category is changed, a new category is created and the update date and time are recorded. Except for the newly created category, the previous features are copied as they are. This allows the user to easily search for past documents and check the contents.

Next, the procedure for finding the characteristics of the document and creating an index in S22 will be described. The creation of the index is performed based on the degree of association between meanings, as shown in FIG.

In step S31, the control unit 11
Corresponds to step S11 in FIG. 4 and step S21 in FIG.
Perform active spreading in the received document. That is, the central activity value of each element in the document is diffused. The central activity value diffusion process will be further described later. The control unit 11 stores, for example, the RAM 14 with the central activation value obtained by the diffusion processing.

In step S32, the control unit 11
Extracts elements whose central activity value exceeds a preset threshold based on the central activity value of each element obtained in step S11. The control unit 11 causes the RAM 14 to store the extracted elements, for example.

In step S33, the control unit 11
Reads the element extracted in step S32 from the RAM 14, for example. The control unit 11 extracts all proper nouns from these elements based on tags and adds them to the index.
To memorize.

Proper nouns do not have a plurality of meanings and have special properties such as not being listed in a dictionary. Therefore, they are treated separately from words other than proper nouns. Proper nouns are identified by tags such as place names, personal names, and organization names as described above.
For example, in the sentence shown in FIG. 3, "Mr. A", "B Association" and "C City" are proper nouns because the tags describe that they are a person name, an organization name, and a place name, respectively. I understand.

In step S34, control unit 11
Extracts a meaning other than a proper noun from the element extracted in step S32 from the RAM 14 and adds it to the index, and stores the result in the RAM 14.

As described above, the index is obtained by finding the features of the document and arranging the features.
That is, for the above-described document in which the internal structure is described by tagging, the central activity value of each vocabulary element is diffused, and vocabulary elements whose central activity value after diffusion is larger than a threshold value are extracted. Add proper nouns or meanings to the index for the extracted vocabulary elements.

It should be noted that the index includes a document address indicating the position of the document stored in the RAM 14 together with the characteristics of the document. Since the index includes words representing the characteristics of the document, it can be used when searching for a desired document. The application of the index to the automatic classification will be further described later.

Here, a specific example of the index will be described.

<Index date = “AAAA / BB / CC”
Time = “DD: EE: FF” Document address = “1234”><Summary> Tax reduction scale, untouched-Prime Minister X's interview </ Summary><Word meaning = "0003" Central activity value = "140.6"> Untouched </ Word><word meaning = “0105” identifier = “X” central activity value =
“67.2”> Prime Minister </ word><Name Identifier = “X” Word Meaning = “6103” Central Activity Value = “150.2”> X Prime Minister </ word><Word Meaning = “5301” Central Activity Value = “120.6” > Determined </ word><word Meaning = “2350” Identifier = “X” Central activity value =
“31.4”> Prime Minister </ word><word meaning = “9582” Central activity value = “182.3”> emphasized </ word><word meaning = “2595” Central activity value = “93.6”> touch </ word><Word meaning = “9472” central activity value = “12.0”> forecasted </ word><word meaning = “4934” central activity value = “46.7”> not touched </ word><word meaning = “0178” Central activity value = "175.7"> Explained </ word><word Meaning = "7248" Identifier = "X" Central activity value =
“130.6”> I </ word><word Meaning = “3684” Identifier = “X” Central activity value =
“121.9”> Prime Minister </ word><word meaning = “1824” Central activity value = “144.4.”> Appealed </ word><word meaning = “7289” Central activity value = “176.8”> Showed </ / Word></index>

In this index, <index> and </ index> indicate the start and end of the index, <date> and <time> indicate the date and time when this index was created, <summary> and </ <Summary> indicates the start and end of the summary of the contents of the index, and <word> and </ word> indicate the start and end of the word. The meaning = “0003” indicates that the meaning is the third of a plurality of meanings. The same applies to other cases. In addition, a central activity value is assigned to each word.

Next, a method of diffusing the central activation value of the element by the activation diffusion performed in step S31 based on the internal structure by tagging will be described. This active diffusion is also performed in step S62 in FIG. 16 described later.

In a document given an internal structure by tagging, a central activity value corresponding to the internal structure by tagging can be given to each element by performing a process called activity diffusion. Active diffusion is a process in which an element associated with an element having a high central activity value is also given a high central activity value. Since the central activity value is determined according to the internal structure by tagging, it can be used for analyzing a document in consideration of the internal structure by tagging.

The active diffusion is executed under the control of the control unit 11 of the document processing apparatus according to a series of steps shown in the flowchart of FIG.

In step S41, for the reference / referenced link and the normal link, the end point activation value of the end point of the link connecting the elements is set to 0. The control unit 11 stores the initial value of the end point activation value thus assigned, for example, in the RAM
14 is stored.

The connection between the elements is as shown in FIG. 10, for example. In this figure, an element _Ei and an element _Ej are shown as a part of the structure of the elements and links constituting the document. The element E _i and the element E _j are the central activation values e _i and e _j
And are connected by a link L _ij . The end point of the link L _ij connected to the element E _i is T _ij , and the end point of the link L _ij connected to the element E _j is T _ji . Element E
_i, in addition to the elements E _j, which is connected by a link L _ij, the link L _ik, L _il and L _im element E _k (not shown) by being connected to the E _l and E _m.
Element E _j, in addition to the elements E _i connected by L _ji is an element E _j viewed from the link L _ij, the link L _uk, L _jq and L not shown by _jr element E _p, the E _q and E _r Each is connected.

[0076] In step S42, the control unit 11 of the document processing apparatus initializes the counter for counting the elements E _i of a document. That is, the count value i of the counter for counting elements is set to one.
This counter will refer to the _first element E1.

In step S43, the control unit 11 of the document processing apparatus executes link processing for calculating a new central activation value for the element referred to by the counter. This link processing will be further described later.

In step S44, the control unit 11 of the document processing device determines whether the calculation of the new central activation value has been completed for all the elements in the document.
Then, when the calculation of the central activation value is completed for all the elements in the document, the control unit 11 sets “YES”.
When the calculation of the new central activation value is not completed for all the elements in the document, “NO” is determined and the process proceeds to step S47.

Specifically, the control section 11 determines whether or not the count value i of the counter has reached the total number of elements included in the document. Then, when the count value i of the counter reaches the total number of elements included in the document, the control unit 11 determines that all elements have been calculated and proceeds to step S45. When the count value i of the counter has not reached the total number of elements included in the document, the control unit 11 determines that the calculation has not been completed for all elements, and proceeds to step S47.

In step S47, the control section 11 of the document processing apparatus increases the count value i of the counter by 1, and sets the count value of the counter to i + 1. Thus, the counter refers to the (i + 1) th element, that is, the next element. Then, the process proceeds to step S4
Returning to 3, the calculation of the endpoint activity value and a series of subsequent steps are executed for the next (i + 1) th element.

More specifically, the control section 11 increases the count value i of the counter for counting elements by one. As a result, the counter refers to the element next to the element for which the central activity value was calculated in step S43.

In step S45, the control unit 11 of the document processing apparatus changes the central activity values of all the elements included in the document, that is, the variation of the newly calculated central activity value from the original central activity value. Calculate the average value for.

The control unit 11 of the document processing device is, for example, R
The original center activity value and the newly calculated center activity value stored in the AM 14 are read for all elements included in the document. The control unit 11 divides the sum of the respective changes of the newly calculated central activity value from the original central activity value by the total number of elements included in the document,
Calculate the average value of the change in the central activity value of all elements. The control unit 11 calculates the average value of the change of the central activity values of all the elements calculated as described above, for example, as
It is stored in AM14.

In step S46, the control unit 11
Determines whether or not the average value of the change in the central activity value of all the elements calculated in step S49 is within a preset threshold value. And the control unit 11
Is "YES" if the change is within the threshold, and the series of steps is terminated. If the change is not within the threshold, the control unit 11 sets the count value i of the counter to 1 in step S42 and executes a series of steps for calculating the central activation value of the document element again in step S42. I do. Step S4 composed of this series of steps
Each time the loop from step 2 to step S44 is repeated, the amount of the change gradually decreases.

Next, the link processing executed in step S43 will be described with reference to the flowchart shown in FIG. This flowchart is a process for one element E _i, during the diffusion process of the central activation value, the link process is performed for all the elements.

[0086] In step S51, the control unit 11 of the document processing apparatus initializes the counter for counting the link elements E _i and one end is connected of a document. That is, the count value j of the counter for counting the links is set to one. That is, this counter refers to the first element L _i1 connected to the element E _i .

In step S52, element E
_In the link L _ij connecting _i and E _j , the control unit 11
Determines whether the link _Lij is a normal link by referring to the tag. The control unit 11 determines whether the link L _ij is a normal link or a reference link. This is determined by referring to the tag of the relation attribute. When the link is a normal link, the control unit 11 proceeds to step S53 as “YES”, and when the link is a reference link, proceeds to step S54 as “NO”.

In step S53, the element E
Processing for calculating a new endpoint activation value of the endpoint T _ij connected to the normal link L _{ij of i} is performed.

Here, it is clear from the determination in step S52 that the link _Lij is a normal link. Point activation values t _ij endpoint T _ij that is normally connected to the link L _ij of the element E _i, of the end-point activation value of the element E _j, all endpoints T _uk connecting link other than the link L _ij, T _jq , T _jr of the end-point activation values t _jp, t _jq, t
and _jr, adds the central activation value e _j of the element E _j of the element E _i is connected by the link L _ij, is determined by dividing the total number of elements included the value obtained by this addition to the document.

The control unit 11 of the document processing apparatus, for example,
From the data stored in the AM 14, necessary end point activation values and center activation values are read. The control unit 11 calculates a new endpoint activity value of the endpoint connected to the normal link as described above for the endpoint activity value and the center activity value that have been read. Then, the control unit 11 stores the calculated end point activation value in the RAM 14, for example.

In step S54, the element E
Processing is performed to calculate the end point activity value of the end point T _ij connected to the reference link of _i .

From the determination in step S52, it is clear that the link _Lij is a reference link.
The new endpoint activation values t _ij of the endpoints T _ij connected to the reference link L _ij of the element E _i are all the endpoints T _jp connected to the links excluding the link L _ij among the endpoint activation values of the element E _j. , T _jq, T _jr of the end-point activation values t _jp, t _jq, t
and _jr, it is obtained by adding the central activation value e _j of the element E _j of the element E _i is connected by a link L _ij.

The control unit 11 of the document processing apparatus, for example,
From the data stored in the AM 14, necessary end point activation values and center activation values are read. The control unit 11 calculates a new endpoint activation value connected to the reference link as described above using the read endpoint activation value and the central activation value. Then, the control unit 11 calculates the endpoint activity value calculated in this manner,
For example, it is stored in the RAM 14.

[0094] processing of the normal link in step S53, and processing of the reference link in step S54, as in a loop extending from the step S52 to step S55, all of which connect to the element E _i that are referenced by the count value i Executed for link L _ij .

In step S55, the element E
_It is determined whether the endpoint activation values have been calculated for all the links connected to _i . If the end point activation values have been calculated for all the links, “YES”
The process proceeds to step S57, and if the end point activation values have not been calculated for all the links, the process proceeds to step S57 as "NO".

In step S56, step S5
At 5 for all links L _ij of the element E _i so that the end-point activation value t _ij is obtained is determined, performing an update of the central activation value e _i of the element E _i.

[0097] the new value that is updated value of the central activation value e _i of the element E _i, the sum e _i of a new end-point activation values of all of the end points of the current central activation value e _i and the element E _i of the element E _i It is obtained by taking '= e _i + _{ t _j '. Here, the prime “′” means a new value.

The control unit 11 of the document processing apparatus, for example,
A necessary endpoint activation value is read from the data stored in the AM 14. The control unit 11 performs the calculation as described above,
Calculating a central activation value e _i of the element E _i. Then, the control unit 11 stores the new central activation value e _i calculated for example RAM 14.

Next, the operation of the document processing apparatus will be described. The document processing apparatus can find not only documents corresponding to the current classification model but also past categories and documents by going back to the date and time desired by the user. That is,
Since the classification of a document depends on the interests of the user, it may change over time. The document processing device
The past classification is also reproduced for the classification that has changed over time.

The document processing apparatus, as shown in FIG.
In 101, the input unit 20 of the document processing apparatus receives an input of a date and time from a user. The input of the date and time by the user is performed, for example, by selecting with a cursor displayed on the display unit 30 in conjunction with the mouse of the input unit 20 in the calendar displayed on the display unit 30. Also,
The input of the date and time by the user may be input by, for example, the keyboard of the input unit 20. The control unit 11 causes the date and time input to the input unit 20 to be stored in, for example, the RAM 14.

In step S102, the control unit 11 of the document processing apparatus reads a classification model, which is a model for classifying documents stored in the RAM 14 into categories, for example. Since the classification model has a history by recording the date and time at the time of update, the classification model at a desired date and time can be read by referring to this date and time and the input date and time.

Here, the updating of the date and time of the classification model will be described. The classification model indicated by A in FIG. 13 includes proper nouns “Mr. A,...” For the categories “sports”, “society”, “computer”, “plant”, “art”, and “event”. "Mr. B, ...", "Company C, Company G,
.., “D type,...”, “Mr. E,...” And “Mr. F, H” have the meanings “baseball (4546), ground (2343),. Labor (3112), Unique (9821), ... "," Mobile (2102), ...
・ ・ ”,“ Sakura 1 (11111), Orange 1 (991
1), “Sakura 2 (11112), Orange 2 (991
2) and “Sakura 3 (11113)” are replaced with the document address “SP” indicating the address of the document corresponding to this classification model.
1, SP2, SP3,... "," SO1, SO2, S
O3, ... "," CO1, CO2, CO3, ...
"," PL1, PL2, PL3, ... "," AR
1, AR2, AR3, ... "and" EV1, EV
2, EV3,..., Respectively.
In this classification model, the update date and time "10:33:27 on January 20, 1999" is recorded. Here, for example, "orange 1" indicates a plant orange, and "orange 2" indicates a color orange. The number in parentheses indicates the document address corresponding to the meaning.

As shown in FIG. 13B, the received date and time is the date and time when the document processing apparatus updated the last update date and time “December 1, 1998
The classification model of “0: 19: 56: 10” is newly added to the document processing apparatus by the reception date and time “10:33 on January 20, 1999.
The document has been updated by receiving the document at “minute 27 seconds.” That is, the classification model shown in B in FIG. 13 has the categories “sports”, “society”, “computer”, “plant”, “ For "art" and "event", proper nouns "Mr. A, ...", "Mr. B, ...",
"Company C, Company G, ...", "Class D, ...", "Mr. E,
... "and" Mr. F "are translated into the meaning" baseball (4546),
Grand (2343), ... "," Labor (311
2), unique (9821), ... "," mobile (21
02), ... "," Sakura 1 (11111), Orange 1
(9911) "," Sakura 2 (11112), Orange 2 (9
912) ”and“ Sakura 3 (11113) ”are assigned to the document addresses“ SP1, SP2, SP ”corresponding to this classification model.
3, ... "," SO1, SO2, SO3, ... ",
"CO1, CO2, CO3, ...", "PL1, PL
2, PL3, ... "," AR1, AR2, AR3, ...
.. "and" EV1, EV2, EV3,... ". The classification model indicated by B in FIG. 13 is a proper noun" H "newly corresponding to the category" event ". "Has been added to the classification model shown in FIG.

Note that the classification model is updated in the following manner.
Not only the increase / decrease of the category and the rearrangement of the category, but also the change to another new category may be possible.

In step S103 of FIG. 12, the control section 11 of the document processing apparatus searches the classification model read out in step S102 for a document corresponding to the classification model. The category of the date and time entered by the user and the classification status of the searched documents are reproduced and displayed on the display unit 30. Then, this series of steps ends.

The search for the document in step S103 is performed as follows.
This is performed based on the classification model read in step S102. FIG. 14 shows the reading of the classification model and the retrieval of a document based on the read other classification model.
This will be described with reference to FIG.

In step S111, the user inputs
In 0, enter the date and time as a reference for searching for a document. In step S112, the control unit 11 extracts a document model based on the date and time input in step S111.
The extraction of the document model will be described with reference to FIG.

[0108] In A of FIG. 15, from time t ₁ to time t ₂ classification model P is, from time t ₂ to time t ₃ classification model Q, are respectively registered. That is, at the date and time t ₂ , the classification model P
Was updated to If you enter the date and time of step S111, for example, X month Y date, in step S112, updating the classification model to the exact date and time t ₀ input in step S111. In this case, the classification model Q that are registered on the date and time t ₀ is extracted.

[0109] In B in FIG. 15, similarly to the A in the figure, from time t ₁ to time t ₂ classification model P is, time t
From ₂ up to date and time t ₃ classification model Q are registered, respectively. In step S111, for example, it is assumed that the date and time of one week ago has been input. In such a case, the model is searched with a predetermined width exactly at the date and time t ₀₁ one week ago. In the figure, during the previous date and time t ₀₁ 6 days after one day and one week prior to the date and time t _01, update date and time t ₂ from the classification model P to the classification model Q is present. Therefore, if you specify one or more days as the predetermined width, updated update time t ₂ classification model Q is selected.

The reference to the classification of the past document in step S112 and the extraction of the past document based on the classification have a benefit when the past classification is considered better.

In step S113, the control unit 11 searches and extracts an unread document when the model extracted in step S112 is updated. The unread document is extracted by referring to the classification model based on the update date and time of the classification model extracted in step S112, the documents included in the classification model are recognized, and the update time of each document is further determined from the index of each document. This is performed by referring to.

In step S114, the control unit 11 displays the unread document extracted in step S113 on the display unit 3.
It is controlled so that it is displayed at 0. The display unit 30 can display a classification model within a predetermined period, for example, in a plurality of windows.

Note that, according to the category of the past classification model reproduced as described above, the document received from the past time to the present can be classified. Classification of a current document by a past classification model is useful if the user determines that the past classification model is better than the current classification model. Also, when it is desired to classify a document to be input by a past classification model, the past classification model can be used.

In the above-described series of steps, the control unit 11 reads out the classification model corresponding to the date and time based on the input of the date and time into the input unit 20, and searches the document by this classification model to go back in time and date. Find out past categories and documents. Therefore, the user can easily search for past documents and check the contents.

Next, a description will be given, with reference to the flowchart shown in FIG. 16, of the calculation of the degree of association between meanings based on the above-mentioned central activity value.

At the first step S61, the control unit 1
1 uses a description of the meaning of a word in the electronic dictionary and creates a network of the meaning by tagging using the electronic dictionary.
That is, a network having a structure by tagging the meanings as described above is created from the reference relation between the meanings of the meanings in the electronic dictionary and the meanings appearing in the description. This corresponds to configuring the internal structure by tagging as shown in FIG. 2 using the highest-order element as an electronic dictionary. The control unit 11 sequentially reads, for example, the meanings and their descriptions stored in the RAM 14 and creates a network. The control unit 14 stores the semantic network created in this manner in, for example, the RAM 14.

The electronic dictionary can be received by the receiving unit 21 of the document processing device from, for example, a communication line. The electronic dictionary is provided by a recording medium 32 such as a CD-ROM. The electronic dictionary provided by the recording medium 32 is reproduced by the recording / reproducing unit 31.

In step S62, step S61
The above-mentioned diffusion processing of the central activity value is performed on the network of the meaning created in the above. By this activation diffusion, the central activation value of each meaning is updated according to the internal structure by tagging provided by the electronic dictionary.

In step S63, step S6
Selects one semantic s _i constituting the network of semantic created in 1, in step S64, suitably changing the initial value of the central activation value e _i vocabulary element E _i corresponding to this meaning s _i , The difference Δ of the central activity value at this time
Calculate e _i .

In step S65, step S6
Corresponds to a change in the initial value of the central activation value e _i of the element E _i at 4 obtains a difference .DELTA.e _j of central activation values e _j of the element E _j corresponding to semantic s _j. In the step S66, the difference Δe _j obtained in step S65 step S
The quotient Δe _j / Δe _i divided by Δe _{i found} in 64 is given by the meaning s _i
Is the degree of association between the meaning and the meaning of the meaning s _j . As the central activity value of a certain meaning is changed in step S64, the central activity value of a related word changes.

In step S67, it is determined whether the calculation of the degree of association between meanings has been completed for all pairs of meanings s _i and s _j . Then, when the calculation of the degree of association between meanings is completed for all the meaning pairs, "YES" is determined, and this series of processing ends. If the calculation of the degree of association between the meanings is not completed for all the pairs of meanings, the process returns to step S63, and the calculation of the degree of association between the meanings is continued for the pair for which the calculation of the degree of meaning is not completed. I do.

The calculated degree of association between meanings is defined between each meaning as shown in FIG. In this meaning table, the degree of association between meanings is normalized and takes a value from 0 to 1. That is, in this table of meaning, "computer", "television", "VT"
R indicates the mutual meaning relation between “computer” and “television”.
The relation between the meaning of "computer" and "VTR" is 0.2
5. The degree of association between meanings of "TV" and "VTR" is 0.60
It is. The control unit 11 causes the RAM 14 to store the word-to-sense association degree created in this way, for example.

In the loop from step S63 to step S67, the control unit stores the necessary value in RAM1
4, and the degree of association between meanings is calculated as described above. The control unit 11 causes the RAM 14 to store the calculated word-to-sense association degree, for example.

Next, a description will be given of document classification using the degree of association between meanings calculated as described above. The document classification using the degree of association between meanings is performed by the GUI shown in FIG.
Is used for document classification in the window 301.

The classification model of each category is extracted based on the central activity value based on the internal structure by tagging. As described above, the control unit 11 of the document processing apparatus extracts an element whose central activity value exceeds a predetermined threshold value in step S32, extracts all proper nouns from this element in step S33, and adds the extracted proper noun to the index. In, meanings other than proper nouns are extracted and added to the index. The classification model is composed of, for example, a category index including an index generated by the above-described procedure.

The automatic classification of documents performed in step S23 in FIG. 7 is performed based on the degree of association between meanings according to a series of procedures shown in the flowchart of FIG. 18 with reference to such a classification model.

At step S71, the control unit 11
Defines the number of intersections of the set of proper nouns included in each category C _i of the classification model and the set of proper nouns of the words extracted from the document and indexed in step S62 as P ( C _i ). Then, the control unit 11 calculates the number P thus calculated.
(C _i ) is stored in the RAM 14, for example.

In step S72, control unit 11
Means the meaning in the index of the document and each category Ci
The sum R (C _i ) of the degrees of association between the meanings is calculated with reference to the table of the meanings of the words in FIG. That is, the control unit 11 calculates the sum R (C _i ) of the degree of association between meanings calculated in step S61 for words other than proper nouns in the classification model. And the control unit 1
1 stores the calculated sum R (C _i ) of the degree of association between meanings in, for example, the RAM 14.

In step S73, item C of the document
_The relevance between document classes, which is the relevance for _i, is defined as Rel (C _i ) = mP (C _i ) + nR (C _i ). Here, the coefficients m and n are constants and are parameters indicating the degree of contribution of the respective values to the degree of association between document classifications. The control unit 11 reads, for example, from the RAM 14 the number P (C _i ) of common sets calculated in step S33 and the sum R (C _i ) of the degree of association between meanings calculated in step S64, and applies the document classification to the above equation. Inter-association degree R
Calculate el (C _i ). The values of these coefficients m and n can be, for example, m = 10 and n = 1. Then, the control unit 11 stores the thus calculated inter-document-class relevance Rel (C _i ) in the RAM 14, for example.

The values of the coefficients m and n can also be estimated using a statistical method. That is, the control unit 11 determines the degree of association between document classifications Re for a plurality of pairs of coefficients m and n.
Given l (C _i ), the above coefficients can be obtained by optimization.

At step S74, control unit 11
Is the degree of association Rel (C _i ) between document classifications for the item C _i
Is the largest of all the items, and when the value of the degree of association between the document classifications exceeds a certain threshold value, the document is classified into the category Ci.
That is, the control unit 11 creates the inter-document class relevance Rel (C _i ) for each of the plurality of categories,
When the maximum degree of association between document classifications Rel (C _i ) exceeds the threshold, the documents are classified into the category C _i . When the maximum degree of inter-document-class relevance Rel (C _i ) does not exceed the threshold value, the document is not classified.

As described above, the procedure of calculating the degree of inter-semantic association between the meanings contained in the document and the classification based on the degree of inter-document association based on the processing processes a document having an internal structure composed of a plurality of elements. This document is classified into a plurality of categories. That is, this procedure calculates the degree of inter-document classification between a document and each category, and determines the category into which the document is classified based on the calculated degree of inter-document classification.

Here, the category for classifying a document is characterized by an index including proper nouns and / or meanings extracted from the document. Then, using such a classification model, the number of proper nouns common to the proper noun included in the classification model of each category and the proper noun extracted from the document is calculated, and the number of proper nouns included in the classification model of each category is calculated. Calculate the sum of the relevance between the meanings of the document with respect to the meaning. Furthermore, the above document is determined by the inter-document relevance based on the number of duplicate proper nouns extracted based on the proper nouns and the meaning relevance included in the document and the sum of the relevance between the meanings other than the proper noun. Decide which category to classify. This degree of association between meanings is determined based on the internal structure of the dictionary by tagging as described above.

Document classification is performed, for example, on items in which the linear combination of the number of common proper nouns and the degree of association between word meanings is maximized, and the degree of association between document and document classes exceeds a predetermined threshold. It is. The coefficient of the linear combination of the number of common proper nouns and the degree of association between meanings can be statistically determined from the magnitude of the association between the document and the category as described above.

Next, the recording medium 32 recorded / reproduced in the recording / reproducing section 31 of the document processing apparatus will be described. The recording medium 32 records a document processing program for processing a document having an internal structure by tagging from a plurality of elements. As the recording medium 32, for example, a floppy disk capable of recording / reproducing information is used.

This recording medium 32 uses a classification model for classifying an electronic document, which is provided with information on an internal structure composed of a plurality of elements, into a plurality of categories, and refers to the update date and time recorded in the classification model. A document processing program for processing the electronic document, the document processing program comprising: an input process for inputting a date and time; and a readout for reading a classification model corresponding to the date and time input in the input process. And a search process for searching for a document based on the classification model read in the read process.

In the present embodiment, an example of a method of tagging a document has been described, but it is needless to say that the present invention is not limited to this tagging method. Also,
In the present embodiment, the document is transmitted from the outside to the receiving unit 21 of the document processing apparatus, but the present invention is not limited to this. For example, the above document is stored in the document processing device R
The data may be written in the OM 13 or read from the recording medium 32 in the recording / reproducing unit 31.

Further, in the above-described embodiment, a mouse is exemplified as a device for selecting a desired element from a document displayed on the display unit 30 of the document processing apparatus.
It goes without saying that the present invention is not limited to this. Other devices such as a tablet and a light pen can be used for inputting elements in the document processing apparatus.

[0139]

As described above, the present invention processes an electronic document by referring to the update date and time of the classification model recorded in the classification model. The document is searched based on the classification model to be searched. Therefore, according to the present invention, not only the past documents but also the classification of the documents into the classification items based on the past classification model can be known. In other words, the present invention also provides past classifications so that it is possible to know the direction of the user's past interests, thereby increasing convenience in use.

Further, according to the present invention, feature information representing features of an electronic document is extracted, and for each of a plurality of classification items constituting a classification model, each document is classified according to the degree of association with the feature information of the electronic document. Classify into items. Therefore, according to the present invention, an electronic document can be classified by feature information based on a past classification model.

[Brief description of the drawings]

FIG. 1 is a block diagram illustrating a configuration of a document processing apparatus to which the present embodiment has been applied.

FIG. 2 is a diagram showing an internal structure by tagging a document.

FIG. 3 is a diagram showing a window displaying an internal structure by tagging a document.

FIG. 4 is a flowchart illustrating an operation of the document processing apparatus to which the embodiment is applied.

FIG. 5 is a diagram showing a GUI for classifying documents.

FIG. 6 is a flowchart for manually classifying documents.

FIG. 7 is a flowchart for automatically classifying documents.

FIG. 8 is a flowchart for finding an index of a document and creating an index.

FIG. 9 is a flowchart showing active diffusion.

FIG. 10 is a diagram illustrating a process of active diffusion.

FIG. 11 is a flowchart of link processing of active spread.

FIG. 12 is a flowchart showing a series of steps for reproducing past classification items and documents by going back to the date and time.

FIG. 13 is a diagram illustrating updating of a date and time of a classification model.

FIG. 14 is a flowchart for extracting an unread document within the range of date and time.

FIG. 15 is a diagram for explaining classification model extraction close to the input date and time.

FIG. 16 is a flowchart of calculation of the degree of association between meanings.

FIG. 17 is a diagram showing a table of the degree of association between meanings.

FIG. 18 is a flowchart of document classification based on the degree of association between document classifications.

[Explanation of symbols]

10 body, 11 control unit, 12 interface,
13 CPU, 20 input unit, 21 receiving unit, 30 display unit, 31 recording / reproducing unit

Claims (10)

[Claims] 1. A document processing method for processing an electronic document by referring to an update date and time of the classification model recorded in the classification model, the method comprising: inputting a date and time; A search step of searching for an electronic document based on a classification model to be processed. 2. The document processing method according to claim 1, wherein when the classification model is updated, the classification model is updated by recording the date and time. 3. A feature information extracting step for extracting feature information representing features of the electronic document from the electronic document; and a feature of the electronic document extracted in the feature information extracting step for a plurality of classification items constituting the classification model. 2. A document processing method according to claim 1, further comprising a classification step of classifying each electronic document into said classification items according to a degree of association with information. 4. A receiving step of receiving a plurality of electronic documents, wherein the characteristic information extracting step extracts, for each electronic document received in the receiving step, characteristic information representing characteristics of the electronic document. The document processing method according to claim 3, wherein 5. The document processing method according to claim 4, wherein the classification model is updated by recording date and time when a new electronic document received in the receiving step is classified. 6. The document processing method according to claim 1, wherein the electronic document is provided with information on an internal structure composed of a plurality of elements. 7. A document processing apparatus for processing an electronic document by referring to an update date and time of the classification model recorded in the classification model, comprising: input means for inputting a date and time; A search unit for searching an electronic document based on a classification model to be processed. 8. The document processing apparatus according to claim 7, wherein the electronic document is provided with information on an internal structure composed of a plurality of elements. 9. A recording medium on which a document processing program for processing an electronic document with reference to an update date and time of the classification model recorded in the classification model is recorded, wherein the document processing program includes an input for inputting a date and time. A recording medium comprising: a process; and a search process for searching for an electronic document based on a classification model corresponding to the date and time input in the input process. 10. The document processing apparatus according to claim 9, wherein the electronic document is provided with information on an internal structure composed of a plurality of elements.