JPH10320409A - Method and device for extracting document information and storage medium storing document extracting process program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To extract a group of document contents as a context since it is hard to extract part of a document or find differences between two documents, if a document is fractionized too much. SOLUTION: Paragraphs are detected from the document contents of one document, the document contents are divided by the paragraphs, and a morpheme analysis is carried out by the paragraphs. Then featured elements are extracted on the basis of the morpheme analytic results (step s1) and a feature table is generated which shows the relation ship between the featured elements and the paragraphs including the featured elements (step s2). On the basis of this feature table, the document is classified by contents as meaningful groups (step s3) and when a content selection indication is received from a user (step s4), the document contents of the paragraph belonging to the selected content are outputted (steps s5 and s6).

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a document information extracting method and apparatus suitable for performing a process for extracting a part of a document content or obtaining a difference between documents. The present invention relates to a storage medium storing an information extraction processing program.

[0002]

2. Description of the Related Art For example, when calculating a difference between two document contents or extracting a part of a certain document content, a document content requested by a user from a database or the like in which document data is stored is stored in an appropriate unit. It is necessary to extract efficiently.

As described above, in order to extract only a specific portion, it is necessary to extract document data by dividing it based on a predetermined rule and display information on the extracted portion.

[0004] On the Internet, when extracting the contents of a part of a document, for example, HTML (H
hyper Text Markup Language)
If the document is written in a document structure description language such as
It is conceivable to perform separation using a delimiter or the like used in the description language, or, for a document written in a natural language, to separate the document by indentation at the beginning of a line or a blank line. By storing information on how such a unit of division is stored in the system in advance, it is possible to separate a document according to the stored contents.

[0005]

However, when various kinds of divisions are stored so as to be applicable to various kinds of documents and applied to all the documents to be processed irrespective of the type and content of the documents, Depending on the document, the cut contents may be too fine, or may be cut where it is better not to cut.

In particular, in a case where a difference between an updated part and information before an update is obtained in online information, if the information is extracted by dividing it into very small pieces, or if it is cut in a place where it cannot be cut, the context of the content may be reduced. Sometimes you don't understand, and you can't understand the meaning.

Accordingly, the present invention detects a segmentation unit (hereinafter referred to as a paragraph) which is segmented by a delimiter or a format from a document, extracts a characteristic element representing the characteristic of the content for each paragraph, and extracts each characteristic. Based on the elements, each paragraph is classified and output as a group of paragraphs that are semantically similar (this is referred to as content in the text). In addition, it is possible to classify and display a plurality of contents, and to output a document content belonging to the content when the user designates the content. Further, it is possible to further re-classify such content in the content. Accordingly, an object of the present invention is to provide a document information extraction method and a document information extraction apparatus which are extremely convenient for performing a process of calculating a difference between two documents or extracting a part of a certain document.

[0008]

According to the first aspect of the information retrieval method of the present invention, a paragraph is detected from the document content of a certain document, and a characteristic element representing the characteristic of the content is extracted for each paragraph. A feature table that represents a relationship between a feature element and a paragraph including the feature element is created, and the document is classified into a plurality of contents for each meaningful unit and output based on the feature table. I have.

According to a second aspect of the present invention, in the first aspect of the present invention, the process of classifying and outputting the document into contents in a meaningful unit is performed based on the feature table. The contents are classified into a plurality of contents and displayed, and when a content selection instruction is received from the user, the contents belonging to the selected contents are output.

According to a third aspect of the present invention, in the first or second aspect of the present invention, the processing for classifying a document into a plurality of contents for each meaningful unit based on the feature table is performed in each paragraph. Based on the element, a plurality of paragraphs having a common characteristic element are made into one unit, and the unit is set as content.

According to a fourth aspect of the present invention, in the second aspect of the invention, the display content when the content is classified and displayed in the semantic unit is at least for each content. This is a representative feature element.

According to a fifth aspect of the present invention, in the second aspect of the present invention, the display content when the content is classified and displayed in the semantic unit is at least for each content. It is assumed that the representative characteristic element and the data related to the paragraph included in the content are used.

[0013] Further, according to the invention of claim 6 in the document information extracting apparatus of the present invention, a document group storage unit for storing a document group;
For a document stored in the document group storage unit, a sentence analysis unit that detects a paragraph from the document content and performs a morphological analysis for each paragraph, and a content for each of the paragraphs based on the analysis result by the sentence analysis unit A feature table creating unit that extracts a feature element representing the feature of the feature and creates a feature table that represents a relationship between the feature element and a paragraph including the feature element; It has a paragraph classifying unit for classifying into a plurality of contents for each unit, a classification result storage unit for storing the contents classified by the paragraph classifying unit, and an output control unit for reading and outputting the contents of the classification result storage unit. It is characterized by:

The contents of the classification result storage section are read and the plurality of contents are displayed.
And a display control unit that, when receiving a content selection instruction from the user, displays the contents of the division unit belonging to the selected content.

According to a seventh aspect of the present invention, in the sixth aspect of the invention, the output control section controls the display of the plurality of contents and, when receiving a content selection instruction from a user, selects the content. Output the contents of the paragraph belonging to the selected content.

According to the invention of claim 8, in the invention of claim 6 or 7, the process of classifying a document into contents of a semantic unit based on the feature table performed by the paragraph classifying unit is performed in each paragraph. Based on the characteristic elements, paragraphs having common characteristic elements are grouped into one, and the group is defined as content.

According to a ninth aspect of the present invention, in the invention of the seventh aspect, the display contents when the content is classified and displayed in the semantic unit are at least for each content. This is a representative feature element.

According to a tenth aspect of the present invention, in the seventh aspect of the present invention, the display contents when the content is classified and displayed in the semantic unit are at least for each content. It is data on representative characteristic elements and paragraphs included in the content.

Further, the invention according to claim 11 in a storage medium storing a document information extraction processing program of the present invention,
A storage medium describing a processing program for performing a document extraction process by a computer, wherein the processing program detects a paragraph from the document content of a certain document, and extracts a characteristic element representing a feature of the content for each paragraph. A feature table that represents a relationship between a feature element and a paragraph including the feature element is created, and the document is classified into a plurality of contents for each meaningful unit and output based on the feature table. I have.

According to a twelfth aspect of the present invention, in the storage medium according to the eleventh aspect, the process of classifying the document into a plurality of contents in a meaningful unit and outputting the content is performed based on the feature table. It is characterized in that it is a process of classifying and displaying a plurality of contents for each summary, and outputting a content of a paragraph belonging to the selected content when receiving a content selection instruction from a user.

According to a thirteenth aspect of the present invention, in the storage medium of the twelfth aspect, the process of classifying a document into a plurality of contents in each meaningful unit based on the characteristic table is performed based on a characteristic element of each paragraph. It is characterized in that a plurality of paragraphs having a common characteristic element are processed into one unit, and the unit is used as a content.

According to a fourteenth aspect of the present invention, the contents displayed by the display processing by the storage medium according to the twelfth aspect (display contents when the contents are classified and displayed in a semantic unit) are each content. Each is characterized by being at least a characteristic element representing the content.

According to a fifteenth aspect of the present invention, the contents displayed by the display processing by the storage medium according to the twelfth aspect (display contents when the contents are classified and displayed in a semantic unit) are each content. Each is characterized by being at least data relating to a characteristic element representing the content and a paragraph included in the content.

According to the present invention, a paragraph is extracted from a document, a characteristic element representing the characteristic of the content is extracted for each paragraph, and the document is classified and output as contents in a meaningful unit. Further, the document is classified and displayed in a plurality of contents for each semantic unit, and when a content selection instruction is received from the user, the contents belonging to the selected content are output.

As described above, according to the present invention, since paragraphs are collectively output as content that is a unit of content, even if the paragraphs are excessively separated by a delimiter or a format, they can be recombined depending on the content. . For this reason, it is possible to set the paragraph division method regardless of the content and type of the document. Also, since paragraph detection can be performed relatively easily, efficient segmentation can be performed.

Further, in the process of classifying a document into contents of a semantic unit, paragraphs having common characteristic elements are classified as contents based on characteristic elements of respective paragraphs.

With this, even if the unit of document separation is a relatively small unit such as a paragraph unit, it is classified as a content composed of a plurality of paragraphs having a semantic unit, so that the document to be processed is finely divided. No more being overkilled.

Further, it is possible to classify and display the document into a plurality of contents, and to output the contents of the document belonging to the contents by designating the contents by the user. When displaying these multiple contents,
Each content consists of at least the characteristic element representing the content or the characteristic element representing the content, and the data related to the paragraphs included in the content. This makes it easier to select desired contents efficiently. In addition, when a content selection instruction is received from the user, by outputting paragraphs belonging to the selected content, unnecessary portions can be omitted for subsequent processing and user browsing, and useful information can be efficiently narrowed down. it can.

Thus, for example, when some processing is performed for each content (for example, when a difference between an updated part of online information and information before update is viewed), a semantic unity is obtained. Since the display is performed in a unit of content composed of a plurality of paragraphs, it is very convenient to understand the context of the content and the overall content.

[0030]

Embodiments of the present invention will be described below with reference to the drawings.

FIG. 1 is a diagram showing an apparatus configuration for realizing the present invention. A document group storage unit 11, a sentence analysis unit 12, a feature table creation unit 13, a paragraph classification unit 14, a classification result storage unit 15, a display It comprises a control unit 16, a display unit 17, and the like.

The document group storage unit 11 stores a large number of documents included in a certain document group as a database.

As a certain one of a large number of documents stored in the document group storage unit 11, for example,
As shown in FIG. 2, if there is a “paper group on artificial intelligence”, as papers belonging to the “paper group on artificial intelligence”, for example, “papers on expert systems”, “papers on natural language processing”, "Articles on Neural Networks" and more.
For example, in a "paper on an expert system", there is a "paper on factory control" and a "paper on the market". I do.

The sentence analysis unit 12 detects paragraphs of each document based on the documents stored in the document group storage unit 11 and performs morphological analysis.

As means for detecting a paragraph, for example, if a blank line exists in a document and it is determined that there is a blank line, the blank line portion is regarded as a paragraph boundary. Considering the beginning or end of a line and comparing it to the beginning of a line, if there is an empty space for one character, consider it as a paragraph boundary, or, if it is a description language such as HTML described above, detect a paragraph symbol to detect a paragraph boundary. There are several methods, such as detecting an error.

FIG. 3 shows an example of the paragraph configuration of one document shown in FIG. 2. For simplicity of description, it is assumed here that the document has six paragraphs. . Hereinafter, these six units will be referred to as paragraphs A1 to A6.

In the paragraph A1, the main morphemes are:
"Expert system", "natural language", "neural network", etc. exist. Paragraph A2 includes "expert system" as a main morpheme, and paragraph A3 includes "expert system" as a main morpheme. , "Natural language", etc., in paragraph A4, "natural language" etc. exists as a main morpheme, in paragraph A5, "neural network" exists as a main morpheme, and in paragraph A6, It is assumed that “natural language”, “neural network”, and the like exist as main morphemes.

The feature table creation unit 13 includes a feature element extraction unit 131, a feature extraction unit 132, and a feature table 133. The characteristic element extracting unit 131 extracts information (referred to as characteristic element) representing the content of each paragraph based on the information extracted by the morphological analysis in the sentence analyzing unit 12.

The feature extraction unit 132 is a feature element extraction unit 131
For example, based on the characteristic elements obtained by
In a document such as that shown by, the number of times which feature element appears in which paragraph and how many times is counted.

Then, a feature table 133 as shown in FIG. 4 is created from the feature elements extracted by the feature element extraction unit 131 and the numbers counted by the feature extraction unit 132.

In the example of the feature table 133 shown in FIG. 4, as the feature elements, "expert system", "natural language", and "neural network" which are the main morphemes appearing in the paragraphs A1 to A6 described above are shown. Have been.

The characteristic element "expert system" appears twice in paragraph A1, once in paragraph A2 and once in paragraph A3, and in paragraphs A4, A5, A
6 indicates that the number of appearances is 0 each. In addition, the characteristic element “natural language” includes paragraphs A1, A3,
Appear once in A6 and A4, respectively, in paragraphs A2 and A5.
Indicates that the number of occurrences of each is zero. The characteristic element "neural network" appears once in each of the paragraphs A1, A5, and A6, and the number of appearances in each of the paragraphs A2, A3, and A4 is zero.

As described above, the feature table 133 indicates which feature element appears in which paragraph and how many times.

The paragraph classifying unit 14 classifies the paragraphs A1 to A6 into a plurality of contents based on the appearance frequency of the characteristic element in each paragraph with reference to the characteristic table 133 having such contents.

For example, the content 1 is a set of paragraphs including “expert system” as a characteristic element. Note that the paragraphs including the feature element of “expert system” are paragraphs A1, A2, and A3, so the number of corresponding paragraphs is “3”.

In the content 2, paragraphs containing "natural language" as a characteristic element are grouped together.
The paragraph containing the characteristic element of “natural language” is A
Since the numbers are 1, A3, A4, and A6, the number of corresponding paragraphs is “4”.

Further, the content 3 is composed of a paragraph including a "neural network" as a characteristic element. The "neural network"
Since the paragraph including the characteristic element is paragraphs A1, A5, and A6, the number of corresponding paragraphs is “3”.

The paragraph classifying unit 14 classifies such characteristic elements and paragraphs having the characteristic elements, and also associates each content with a paragraph. For example, paragraphs belonging to content 1 are paragraphs A1, A2, A3
And the paragraphs belonging to the content 2 are the paragraphs A1, A3,
A4 and A6, and the paragraph belonging to content 3 is paragraph A
The respective contents such as 1, A5 and A6 are associated with paragraphs belonging to the contents.

The classification result as described above is stored in the classification result storage 1
5 is stored.

The output control section 16 reads out the contents of the classification result storage section 15 and outputs it as a classification result. At this time, control is performed to display the classification result on the display unit 17, and when a search result display instruction is given from the user, the classification is performed based on the contents of the classification result storage unit 15 and the contents of the document group storage unit 11. It is also possible to configure the data as and output the data.

FIG. 5 shows an example of the classification result displayed on the display section 17. In this example, only three contents, namely, contents 1, contents 2, and contents 3 are shown. Although only contents 1 to 3 are shown on the screen, in reality, there are usually many more contents. In such a case, for example, about ten contents are displayed on one screen in consideration of the visibility of the screen. In addition, the number of contents that can be displayed on one screen can be set to an appropriate number, and when the number of contents is large,
It is also possible to divide into about ten pages and display them by switching pages.

In the example of the display contents in FIG. 5, the characteristic element of the content 1 is “expert system”,
The number of paragraphs is "3", the characteristic element of the content 2 is "natural language", the number of paragraphs is "4", and the characteristic element of the content 3 is "neural network".
And the number of paragraphs is displayed as “3”.

As described above, for each content, the characteristic element representing the content and the number of paragraphs including the characteristic element are displayed. The display unit 17 has a user instruction unit 21 for inputting a user instruction such as “display result”.
Is displayed.

The display contents shown in FIG. 5 include, for example, “content 1”, “expert system”,
The number indicating the content, its characteristic element, and the number of paragraphs, such as “3”, are not limited to this. For example, the corresponding paragraph number may be added to the display content. it can. As an example, “content 1”, “expert system”, and “3” are displayed, and as the corresponding paragraph, “paragraphs A1, A2, A
For example, the content indicating a paragraph (for example, a paragraph number or the like) such as “3” may be displayed.

The user looks at such display contents and finds that the information requested by the user is, for example, the content 1
Is determined to be related to the content (characteristic element is “expert system”), the user clicks the row portion R1 of the content 1 with a mouse or the like, and then clicks the user instruction unit 21 of “display result”. .

As a result, the output control unit 16 reads the paragraph belonging to the selected content 1 from the document group storage unit 11 and outputs the content of the content. In this example, the number of paragraphs belonging to the selected content 1 is “3”, and the contents are paragraphs A1, A2, and A3. Therefore, when the output control unit 16 receives a result display request from the user, A1, A belonging to the selected content
The contents of A2, A3 are read from the document group storage unit 11 and the contents are displayed as shown in FIG. 6. FIG. 6 shows an example in which the respective document contents of paragraphs A1, A2, A3 belonging to the selected content 1 are displayed. This example shows an example in which all the document contents of three paragraphs A1, A2, and A3 are displayed on one screen. However, if the contents cannot be displayed on one screen, the page is turned. So that you can see it.

FIG. 7 is a flowchart showing the processing procedure of this embodiment described above. In FIG. 7, first, a paragraph is detected from a document, a morphological analysis is performed, a characteristic element is extracted based on the result of the morphological analysis, and a characteristic as shown in FIG. 4 is obtained based on the extracted characteristic element. A table 133 is created (step s1). The above processing is performed by the sentence analysis unit 12 and the feature table creation unit 13.
Next, the paragraph classifying unit 14 classifies each paragraph into contents with reference to the feature table 133 (step s).
3) The classification result is stored in the classification result storage unit 15 and displayed on the display unit 17 (step s4). An example of the classification result displayed at this time is, for example,
The contents are as shown in FIG.

When the user looks at the display and selects the desired content and inputs "display result" from the user instruction section 21, the output control section 16 receives the user instruction (step s4). ), Output data is constructed from the contents of the document group storage unit 11 and the classification result storage unit 15 (step s5), and the data is output (step s6). An example of this output data is, for example, the content as shown in FIG.

As described above, in this embodiment, first, a paragraph is extracted from a document, characteristic elements representing the characteristics of the contents are extracted for each paragraph, and each paragraph is extracted based on each characteristic element. Is regarded as one unit having the same characteristic element, and is classified as content.

As described above, since the document can be separated regardless of the content and type of the document, any document can be dealt with, and the document content of the separated paragraph can be directly converted into one paragraph unit. Instead of using them, based on the characteristic elements of each paragraph, if they have common characteristic elements, the paragraphs are considered as a semantic unit, and the unit is extracted as one content. .

Therefore, even if the unit of document separation performed by the document analysis unit 2 is a paragraph, the output content is not a paragraph unit, but a content unit composed of a plurality of paragraphs having a meaningful unit. Is output as
One output unit is not over-subdivided.

Thus, as described above, in the case where the difference between the updated part and the information before the update is viewed in the online information, the content is output as a content composed of a plurality of paragraphs having a meaningful unity. By doing so, it is extremely convenient to understand the context of the content and the overall content.

Although the above-described embodiment is an example of a preferred embodiment of the present invention, the present invention is not limited to this, and various modifications may be made without departing from the spirit of the present invention. It will be feasible. Further, a processing program for performing the processing of the present invention can be stored in a storage medium such as a floppy disk, an optical disk, or a hard disk, and the present invention also includes such a storage medium. A format for obtaining data may be used.

[0064]

As described above, according to the present invention, a paragraph is extracted from a certain document, a characteristic element is extracted for each paragraph, and the document is classified and output as a semantic unit. . In addition, the content is classified into a plurality of contents and displayed, and upon receiving a content selection instruction from the user, the contents of the document belonging to the selected content are output. It does not matter what type of document or paragraph you use, and can handle any document. Then, instead of using the paragraph as it is, it is output in a content unit composed of a plurality of paragraphs having a semantic unit, so that one output unit is not excessively divided.

In the process of classifying a document into contents of a semantic unit, paragraphs having common characteristic elements are classified as contents based on characteristic elements of respective paragraphs. Even when the paragraph is divided too finely by applying the paragraph division method corresponding to the type of document, the document to be processed is output in the content unit consisting of multiple paragraphs having a semantic unit. Is not extracted in a subdivided state so that the meaning is not understood.

Further, when a document is classified and displayed as contents in a meaningful unit, the display contents are as follows:
Each content consists of at least a characteristic element representing the content, or a characteristic element representing the content, and data related to paragraphs included in the content. This makes it easy to select desired contents efficiently.

When a content selection instruction is received from the user, the content of the paragraph belonging to the selected content is output, so that the output content is not a document content of one paragraph but a semantic unit. Is output in a plurality of paragraphs having Accordingly, for example, when some processing is performed in content units (for example, when a difference between an updated part and information before update is viewed in online information), a plurality of semantic units are set. Since the content is output in units of content composed of paragraphs, it is extremely convenient for understanding the context of the content and the overall content.

[Brief description of the drawings]

FIG. 1 is a block diagram showing a configuration of a document information extraction device according to an embodiment of the present invention.

FIG. 2 is a diagram showing an example of a document group used in the embodiment of the present invention.

FIG. 3 is a diagram showing an example of a paragraph configuration of a certain document used in the embodiment of the present invention.

FIG. 4 is a diagram showing an example of a feature table according to the embodiment of the present invention.

FIG. 5 is a diagram showing a display example of a plurality of contents according to the embodiment of the present invention.

FIG. 6 is an exemplary view showing an example of outputting paragraph contents belonging to a content selected from the plurality of contents shown in FIG. 4;

FIG. 7 is a flowchart illustrating a processing procedure according to the embodiment of the invention.

[Explanation of symbols]

 Reference Signs List 11 Document group storage unit 12 Sentence analysis unit 13 Feature table creation unit 14 Paragraph classification unit 15 Classification result storage unit 16 Output control unit 17 Display unit 21 User instruction unit of “Result display” 131 Feature element extraction unit 132 Feature extraction unit 133 Features Table A1 to A6 paragraph

Claims (15)

[Claims] Claims 1. A paragraph is detected from the contents of a document,
For each of the paragraphs, a feature element representing a feature of the content is extracted, a feature table representing a relationship between the feature element and a paragraph including the feature element is created, and based on the feature table,
A document information extracting method, wherein the document is classified and output into a plurality of contents for each semantic unit. 2. A process of classifying and outputting the document into a plurality of contents of each semantic unit, and displaying and classifying the document into a plurality of contents of each semantic unit based on the feature table. 2. The document information extracting method according to claim 1, wherein when receiving a content selection instruction from a user, the contents of a paragraph belonging to the selected content are output. 3. A process for classifying a document into a plurality of contents in each meaningful unit based on the characteristic table, wherein a plurality of paragraphs having a common characteristic element are grouped into one based on the characteristic element of each paragraph. 3. The document information extracting method according to claim 1, wherein the set is a content. 4. The display content when the content is classified and displayed in each semantic unit, and at least, for each content, at least a characteristic element representing the content. Document information extraction method described. 5. The display contents when the contents are classified and displayed in the semantic unit are, for each content, at least a characteristic element representing the content and data on a paragraph included in the content. 3. The document information extracting method according to claim 2, wherein: 6. A document group storage unit for storing a document group, and a sentence analysis unit for detecting a paragraph from the document content of the document stored in the document group storage unit and performing morphological analysis for each paragraph. A feature table creation unit that extracts a feature element representing a feature of the content for each paragraph from the analysis result by the sentence analysis unit and creates a feature table representing a relationship between the feature element and a paragraph including the feature element A paragraph classifying unit that classifies the document into a plurality of contents in each of meaningful units based on the contents of the feature table; a classification result storage unit that stores the contents classified by the paragraph classifying unit; A document information extraction device, comprising: an output control unit that reads and outputs the contents of a result storage unit. 7. The output control unit controls the display of the plurality of contents and, when receiving a content selection instruction from a user, outputs the contents of a paragraph belonging to the selected content. The document information extracting device according to claim 6, wherein 8. A process of classifying a document into contents of each semantic unit based on a feature table performed by the paragraph classifying unit, wherein a plurality of paragraphs having a common feature element are classified based on a feature element of each paragraph. 8. The document information extracting device according to claim 6, wherein the document information is extracted as one unit, and the unit is set as a content. 9. The display content when the content is classified and displayed in each semantic unit, and at least a characteristic element representative of the content is displayed for each content. Document information extraction device. 10. The display contents when the contents are classified and displayed in each of the semantic units are, for each content, at least a characteristic element representing the content and data on a paragraph included in the content. 8. The document information extracting device according to claim 7, wherein: 11. A storage medium in which a processing program for performing a document extraction process by a computer is described, wherein the processing program detects a paragraph from the document content of a certain document, and indicates a characteristic element representing a feature of the content for each paragraph. To create a feature table representing the relationship between the feature element and the paragraph containing the feature element, and based on the feature table, classify the document into a plurality of contents for each semantic unit and output it A storage medium storing a document information extraction processing program characterized by performing the following. 12. The process of classifying and outputting the document into a plurality of contents in each of the semantic units, wherein the document is classified into a plurality of contents in each of the semantic units based on the feature table and displayed. And outputting a content of a paragraph belonging to the selected content when receiving a content selection instruction from a user.
A storage medium storing a document information extraction processing program described in Item No. 0. 13. A process of classifying a document into a plurality of contents for each semantic unit based on the characteristic table, wherein a plurality of paragraphs having a common characteristic element are grouped into one based on the characteristic element of each paragraph. 2. The process according to claim 1, wherein the process is a process in which the content is set as a content.
A storage medium storing the document information extraction processing program according to 1 or 12. 14. The display content when the content is classified and displayed in each semantic unit, and is at least a characteristic element representing the content for each content. A storage medium storing the described document information extraction processing program. 15. The display contents when the contents are classified and displayed in the semantic unit are, for each content, at least a characteristic element representing the content and data on a paragraph included in the content. 13. A storage medium storing the document information extraction processing program according to claim 12.