Index preparing device and index utilizing device

Description: TECHNICAL FIELD [Detailed description of the invention]

【0001】

[Industrial application field]

The present invention relates to an index creation device and an index utilization device that automatically generate an index for searching a corresponding part from a keyword for one document.

【0002】

[Conventional technology]

In the case of technical books, etc., for important keywords, a page containing related matters, chapter / section numbers, etc. should be grouped together, and an index arranged in alphabetical order or alphabetical order should be recorded at the end of the book. Is commonly done. With such an index, readers can quickly open the section that corresponds to what they want to know. Indexes are generally created by humans, but they have the following problems.

【0003】

1. Keyword creation is troublesome. 2. There is a possibility that important keywords may not be recorded. 3. It is troublesome to associate keywords with corresponding parts.

【0004】

[Problems to be Solved by the Invention]

In order to save the trouble of creating an index by a human being, for example, the following method is used as a conventional method.

【0005】

1. For books, enclose the keywords you want to index in the document with \ index {and} . 2. Layout the document and create a correspondence table between the keywords and page numbers enclosed in \ index {and} . 3. Generate an index from the correspondence table. That is, it is still necessary for humans to automatically associate keywords with corresponding parts, determine keywords to be indexed, and mark them.

【0006】

In this case, the problem that it is troublesome to associate the keyword with the corresponding part has been solved, but 1. Creating keywords is troublesome. 2. There is a risk of omission of recording of important keywords. That point has not been resolved. In this way, the conventional automatic keyword extraction technology is to search for a document containing the items that the reader wants to know from a plurality of documents, and is an index for searching for the items that the reader wants to know in one document. Cannot be generated.

【0007】

[Means for solving problems]

In the invention according to claim 1, a keyword extraction means for extracting a keyword from a structured document, a keyword storage means for storing the extracted keyword and a position on the document in association with each other, and an accessible format for the position on the document. An index creation device is configured by a position conversion means for converting to and an index generation means for generating an index from a set of the keyword and a position on the document in the accessible format.

【0008】

In the invention according to claim 2, in the invention according to claim 1, an extraction rule storage means for converting a keyword extraction rule according to a document element is provided.

【0009】

In the invention according to claim 3, in the invention according to claim 1, an index element storage means for unconditionally indexing the contents of a specific document element is provided.

【0010】

In the invention according to claim 4, in the invention according to claim 1, a keyword selection means for human beings to determine whether or not the extracted keyword should be an index is provided.

【0011】

In the invention according to claim 5, a keyword extraction means for extracting keywords and importance from a structured document, and a keyword storage means for storing the extracted keywords, the importance, and a position on a document in association with each other. A position conversion means for converting a position on a document into an accessible format, an index generation means for generating an index from a pair of the keyword and a position on the document in the accessible format, and an index creation device are configured.

【0012】

In the invention according to claim 6, the keyword designation means for designating the keyword of interest from the documents having the keyword, the position on the document, and the importance as an index, and the position on the document corresponding to the keyword are of importance. A means for presenting related parts to be presented in descending order and a device for using an index were constructed.

【0013】

In the invention according to claim 7, a keyword designation means for designating a keyword of interest from a document having a keyword, a position on a document, and a importance as an index, and a lower limit of importance for presenting a related part are designated. A means for designating the lowest importance, a means for presenting related parts that present only those having a importance equal to or higher than the lowest importance at the position on the document corresponding to the keyword, and a more index utilization device are configured.

【0014】

In the invention according to claim 8, a keyword designation means for designating a keyword of interest from a document having a keyword, a position on a document, and an importance as an index, and the number of presentation-related points for one keyword are used. A means for specifying the maximum number of related parts for designating an upper limit, a means for presenting related parts in which the positions on the document corresponding to the keyword are presented in the order of the number of the maximum number of related parts or less, and a more index-using device are configured.

【0015】

[Action]

In the invention according to claim 1, it is possible to automatically generate an index for searching for matters that the reader wants to know in one document.

【0016】

In the invention described in claim 2, keywords should be broadly taken from important document elements such as "title", and keywords should not be taken from document elements that are not suitable for taking keywords such as "citation" and "example". It becomes possible to do.

【0017】

In the invention according to claim 3, it is possible to specify a word to be included in the index.

【0018】

The invention of claim 4 allows humans to reject words that are not suitable for the index.

【0019】

In the invention according to claim 5, it is possible to automatically add a numerical value expressing the magnitude of the relationship called "importance" to the correspondence relationship between the keyword and the related part.

【0020】

In the invention according to claim 6, it is possible to present the positions of the keywords on the document in descending order of importance.

【0021】

In the invention according to claim 7, it is possible to present only those having a degree of importance equal to or higher than the minimum importance specified by the position of the keyword in the corresponding document.

【0022】

In the invention according to claim 8, it is possible to present as many positions in the document corresponding to the keywords as the number of the maximum number of related points or less.

【0023】

[Example]

An embodiment of the invention according to claims 1 to 4 will be described with reference to FIGS. 1 and 2. FIG. 1 shows the overall configuration of the indexing apparatus. This device has a keyword extraction means 1 that extracts keywords from a structured document (described later), an extraction rule storage means 2 that converts keyword extraction rules according to document elements, and a human being can determine whether the extracted keywords should be used as an index. Keyword selection means 3 for determination, index element storage means 4 that unconditionally indexes the contents of a specific document element, and keyword storage means 5 that stores the extracted keywords and their positions on the document in association with each other. It consists of a position conversion means 6 that converts a position on a document into an accessible format, and an index generation means 7 that generates an index from a pair of a keyword and a position on a document in an accessible format.

【0024】

Here, the "document" which is the object of the present invention will be described. The document here means a "structured document". This structured document is a document in which the content is expressed so that it can be processed as a tree structure of document elements. As an example of this structured document, for example, there is a document having the contents listed in Table 1.

【0025】

[table 1]

<img file="JPH06348756A_D0001.tif" />【0026】

Here, the character string between "&lt;" and "&gt;" is the mark indicating the start of the document element, and the character string between "&lt;/" and "&gt;" is the mark indicating the end of the document element. .. The first string in this mark is the name of the element, and the second string is the ID of the element. Since the ID is guaranteed to be unique within the document, specifying the ID will identify one document element.

【0027】

Next, the specific configuration contents of each of the above-mentioned means will be described. First, the keyword extraction means 1 will be described. This means 1 refers to a process of automatically extracting a keyword to be indexed from a part other than the mark of the document content. In this case, an example of the processing procedure is as follows.

【0028】

1. Skip the mark of the document content. 2. Perform morphological analysis of the character string and extract all nouns. 3. Reject the nouns listed in the unnecessary word dictionary created in advance. In addition, in such a process, the conventional keyword automatic extraction technique can be applied as it is.

【0029】

The extraction rule storage means 2 will be described. This means 2 is for changing the keyword extraction process according to the document element. For example, there are processing contents as shown in Table 1.

【0030】

[Table 2]

<img file="JPH06348756A_D0002.tif" />【0031】

The contents of Table 1 express the following.

【0032】

* Do not use an unnecessary word dictionary when extracting keywords from the contents of document elements as "titles". * When extracting a keyword from the content of a document element called "paragraph", normal processing is performed. * Do not extract keywords from the content of the document element "citation".

【0033】

Keyword selection means 3 will be described. In general, not all automatically extracted keywords are suitable as indexes. Noise is inevitably picked up in the automatic extraction process. It doesn't really hurt if the index contains words that the reader doesn't use, but it doesn't hurt to have more indexes and more time to search. Therefore, this means 3 is provided to reject words that humans think are inappropriate as an index from the automatically extracted keywords. The specific processing procedure of this means 3 is as follows.

【0034】

1. Arrange and display keywords in Aiueo order or ABC order. 2. Humans look at the list display and instruct unnecessary keywords.

【0035】

The index element storage means 4 will be described. Automatically extracted keywords are not always sufficient as an index. In the automatic extraction process, only keywords that appear as the contents of the document can be picked up. As long as the index is created by humans, even words that do not appear in the document can be recorded as an index. Therefore, this means 4 is provided so that keywords that humans think are appropriate can be added. The specific processing procedure of this means 4 is as follows.

【0036】

1. Ask humans for the presence or absence of keywords to be added for each "chapter" or "section". 2. Ask them to enter any additional keywords. 3. Memorize the set of "chapter" or "section" that corresponds to the additional keyword.

【0037】

Keyword storage means 5 will be described. The present means 5 refers to a process of memorizing a set of a keyword and a position in a corresponding document. The position in the document is represented by the element ID uniquely assigned to the document element. The specific processing procedure of the present means 5 corresponding to the above-mentioned structured document is as follows.

【0038】

[Table 3]

<img file="JPH06348756A_D0003.tif" />【0039】

The keywords and positions in Table 3 express the following contents.

【0040】

* The keyword "information retrieval" corresponds to the element attribute with the ID T1. * The keyword "information retrieval" corresponds to the element attribute with the ID P1. * The keyword "keyword" corresponds to the element attribute with the ID T1.

【0041】

The position changing means 6 will be described. The present means 6 performs layout processing and converts the element ID in the keyword storage means 5 into a page number or a chapter / section number.

【0042】

The index generation means 7 will be described. The present means 7 converts the final data into a form that is easy to use. The specific processing procedure of this means 7 is as follows.

【0043】

* Create a correspondence table between keywords and page numbers. * Create a correspondence table between keywords and chapter / section numbers. * Create hyperlinks between keywords and corresponding elements.

【0044】

Next, an operation example of the present device provided with various means as described above will be described based on the flow of FIGS. 2 (a) to 2 (e). First, keywords are automatically extracted from a document (structured document) using the keyword extraction means 1 (a). At this time, processing is performed according to the contents of the extraction rule storage means 2. Next, the extracted keywords are presented to the user, and keywords unnecessary for the index are selected by the keyword selection means 3. This selected keyword is rejected (b). Next, inquire about the existence of the keyword that the user wants to add, and if there is that keyword, also inquire which chapter or section to associate with (c). Next, layout processing is performed by the position conversion means 6, and the element ID in the keyword storage means 5 is converted into a page number or a chapter / section number (d). Next, the index generation means 7 converts the combination of the keyword and the position information into a form that is easy to use, and performs index generation (e).

【0045】

As described above, by configuring the indexing device as shown in Fig. 1, it is possible to automatically generate an index to search for what the reader wants to know in one document. This eliminates the need for authors and editors to do the tedious indexing work, and also prevents them from forgetting to index important words. In addition, it is possible to take a wide range of keywords from important document elements such as "title" and not to take keywords from document elements that are not suitable for taking keywords such as "citation" and "example". As a result, it is possible to suppress the increase of meaningless index (noise) and prevent the important keywords from being received. Furthermore, since the words to be included in the index can be specified, important words that do not appear in the contents of the document can be included in the index. Furthermore, since humans can reject words that are not suitable for the index, meaningless or unimportant words can be excluded from the index.

【0046】

Next, an embodiment of the invention according to claim 5 will be described with reference to FIGS. 3 and 4. In the indexing apparatus that operates based on FIG. 2 described in the inventions of claims 1 to 4 described above, the following problems newly arise. That is, if the automatically extracted keywords are indexed as they are, the parts that are not appropriate as related parts are also indexed and recorded. In addition, it is troublesome for humans to select the automatically extracted keywords, and there is a risk that the selection will be subjective and the relevant parts that should be originally necessary will be dropped.

【0047】

Therefore, in this embodiment, the index creation device is configured as shown in FIG. That is, the present device includes a keyword extraction means 8 that extracts keywords and importance from a structured document, a keyword storage means 9 that stores these extracted keywords, importance, and position on a document in association with each other, and a document. It consists of a position conversion means 10 that converts the upper position into an accessible format, and an index generation means 11 that generates an index from a pair of a keyword and a position on a document in an accessible format. Hereinafter, the specific configuration contents of each of these means will be described.

【0048】

First, the keyword extraction means 8 will be described. The present means 8 refers to a process of automatically extracting a keyword to be indexed from a part other than the mark of the document content. For such processing, the conventional keyword extraction technique can be applied as it is. The specific processing procedure of this means 8 is as follows.

【0049】

1. Skip the mark of the document content. 2. Perform morphological analysis of the character string and extract all nouns. 3. Reject the nouns listed in the unnecessary word dictionary created in advance. 4. Record the number of the same keywords that appear in the same document element. The value obtained by dividing the number of each keyword by the maximum number is taken as the importance of a certain keyword for a certain document element.

【0050】

Keyword storage means 9 will be described. This means 9 stores a set of a keyword and a corresponding position in a document and an importance. The position in the document is represented by the element ID uniquely assigned to the document element. The specific processing procedure of the present means 9 corresponding to the above-mentioned structured document is as follows.

【0051】

[Table 4]

<img file="JPH06348756A_D0004.tif" />【0052】

The keywords, positions, and importance in Table 4 express the following contents.

【0053】

* The keyword "information retrieval" corresponds to the element attribute with the ID T1 with a severity of 0.1. * The keyword "information retrieval" corresponds to the element attribute with the ID P1 with a importance of 0.3. * The keyword "keyword" corresponds to the element attribute with the ID T1 with a importance of 0.2.

【0054】

The position changing means 10 will be described. The present means 10 converts the element ID in the keyword storage means 9 into a page number, a chapter / section number, or the like.

【0055】

The index generation means 11 will be described. The present means 11 is a process of converting the final data into a form that is easy to use. The specific processing procedure is as follows.

【0056】

* Create a correspondence table between keywords, page numbers, and importance. * Create a correspondence table between keywords, chapter / section numbers, and importance. * Create hyperlinks between keywords, corresponding elements, and importance.

【0057】

Next, an operation example of the present device provided with various means as described above will be described based on the flow of FIGS. 4 (a) to 4 (c). First, keywords are automatically extracted from the document using the keyword extraction means 8 (a). The set of the position and the importance of this keyword in the corresponding document is stored in the keyword storage means 9. Next, the position conversion means 10 is used to perform layout processing, and the element ID in the keyword storage means 9 is converted into a page number or a chapter / section number (b). Next, the index generation means 11 converts the combination of the keyword, the position information, and the important into an easy-to-use form, and performs index generation (c).

【0058】

As described above, when a keyword is automatically extracted, a numerical value that expresses the importance of the keyword called "importance" is automatically added, and a plurality of associations associated with one keyword are used using the numerical value. An order relationship is set between the locations, and the order of presentation to the user and the number of presentations are adjusted based on the order relationship. Therefore, from such a thing, even if the automatically extracted keyword is indexed as it is, it is possible to construct an index so that a reference to an inappropriate part as a related part is unlikely to occur.

【0059】

Next, an embodiment of the invention according to claims 6 to 8 will be described with reference to FIGS. 5 and 6. FIG. 5 shows the configuration of this index utilization device. This device presents the keyword designation means 12 that specifies the keyword of interest from the documents that have the keyword, the position on the document, and the importance as an index, and the position on the document corresponding to the keyword in descending order of importance. Or, only the position on the document corresponding to the keyword that has the importance equal to or higher than the minimum importance is presented, or the position on the document corresponding to the keyword is the number that is more important than the maximum number of related parts. Related part presentation means 13 to be presented from, minimum importance specification means 14 to specify the lower limit of importance to present related parts, and maximum number of related parts to specify the upper limit of the number of presentation related parts for one keyword. It consists of designated means 15. Hereinafter, the specific configuration contents of each of these means will be described.

【0060】

First, the keyword designation means 12 will be described. The means 12 is a process of designating a keyword corresponding to the item that the reader wants to find. The processing procedure of the specific specification method is as follows.

【0061】

* Enter the keyword you came up with from the keyboard. * Post a list of keyboards and mark the keywords you are looking for.

【0062】

The minimum importance designation means 14 will be described. This means 14 specifies the lowest importance of presenting among the related parts corresponding to the input keyword. Related parts with a lower importance than this number will not be presented.

【0063】

The means for designating the maximum number of related points 15 will be described. The present means 15 specifies the maximum number of related parts to be presented for one keyword.

【0064】

The related part presentation means 13 will be described. The present means 13 presents a related part corresponding to one keyword. As a specific presentation method, for example, there are three presentation methods as described below. As the first presentation method, there is a sequential presentation method (corresponding to the invention according to claim 6).

【0065】

1. Present the most important relevant points. 2. The user instructed to display the following related parts al, then presents a high importance related points. 3. Repeat this procedure until the user tells you to finish.

【0066】

As the second presentation method, there is a sufficient presentation method (corresponding to the invention according to claim 7).

【0067】

1. Present the most important relevant points. 2. When the user instructs to display the next related part, the next most important related part is presented. 3. This procedure is repeated until there are no related parts having an importance equal to or higher than the minimum importance specified by the minimum importance designation means 14, or the user instructs the end.

【0068】

As the third presentation method, there is a simultaneous presentation method (corresponding to the invention according to claim 8).

【0069】

1. Maximum number of related parts The number of related parts specified by the means 15 is taken out in order from the one with the highest importance, and displayed in separate windows at the same time.

【0070】

Next, an operation example of the present device provided with various means as described above will be described based on the flow of FIGS. 6 (a) to 6 (e). First, the search keyword is input by the keyword designation means 12 (a). Next, the related part presenting means 13 presents the related part corresponding to the input keyword (b). In this case, the presentation modes include sequential presentation, shortage presentation, and simultaneous presentation.

【0071】

If the presentation mode is sequential presentation, the following processing is performed (c). 1. Present the most important relevant points. 2. When the user instructs to display the next related part, the next most important related part is presented. 3. Repeat this procedure until the user tells you to finish.

【0072】

If the presentation mode is full presentation, the following processing is performed (d). 1. Present the most important relevant points. 2. When the user instructs to display the next related part, the next most important related part is presented. 3. This procedure is repeated until there are no related parts having an importance equal to or higher than the minimum importance specified by the minimum importance designation means 14, or the user instructs the end.

【0073】

If the presentation mode is simultaneous presentation, the following processing is performed (e). 1. Maximum number of related parts The number of related parts specified by the means 15 is taken out in order from the one with the highest importance, and displayed in separate windows at the same time.

【0074】

As described above, since the positions of the keywords on the document can be presented in descending order of the important part, the target information can be quickly accessed without repeatedly looking at different places in the document. In addition, since it is possible to present only those having an importance equal to or higher than the minimum importance specified by the position on the document corresponding to the keyword, it is possible to eliminate access to almost unrelated parts. Furthermore, since the number of positions on the document corresponding to the keyword can be presented in the number less than the maximum number of related parts, simultaneous display of related parts on a narrow screen and efficient display even when accessing through a slow communication path can be performed. It can be carried out.

【0075】

[Effect of the invention]

The invention according to claim 1 has a keyword extraction means for extracting a keyword from a structured document, a keyword storage means for storing the extracted keyword and a position on the document in association with each other, and a form in which the position on the document can be accessed. Since the index creation device is configured with the position conversion means for converting to, the index generation means for generating the index from the pair of the keyword and the position on the document in the accessible format, and the index creation device, the reader knows in one document. It can automatically generate an index to find what you want, eliminating the need for authors and editors to do the tedious indexing work and forgetting to index important words. It can be done.

【0076】

In the invention of claim 2, in the invention of claim 1, since the extraction rule storage means for converting the keyword extraction rule by the document element is provided, the keyword is widely taken from the important document element such as "title". It is possible to prevent keywords from being taken from document elements that are not suitable for taking keywords such as "quote" and "example", which suppresses the increase of meaningless index (noise) and important keywords. Can be prevented from leaking.

【0077】

The invention according to claim 3 is provided with an index element storage means for unconditionally indexing the contents of a specific document element in the invention according to claim 1, so that words to be included in the index can be specified. , It is possible to index important words that do not appear in the contents of the document.

【0078】

The invention according to claim 4 provides a keyword selection means for a human to determine whether or not the extracted keyword should be an index in the invention according to claim 1, and thus rejects a word that is not suitable for an index by a human. This allows you to remove meaningless or insignificant words from the index.

【0079】

The invention according to claim 5 is a keyword extraction means for extracting keywords and importance from a structured document, and a keyword storage means for storing the extracted keywords, the importance, and a position on a document in association with each other. Since the index creation device is configured with the position conversion means for converting the position on the document into the accessible format, the index generation means for generating the index from the pair of the keyword and the position on the document in the accessible format, and so on. A numerical value that expresses the magnitude of the relationship called "importance" can be automatically added to the correspondence between the keyword and the related part, and this numerical value is used to associate with one keyword. An order relationship can be set between a plurality of related points, and the number of presentations to the user can be adjusted based on the order relationship.

【0080】

In the invention described in claim 6, the keyword designation means for designating the keyword of interest from the documents having the keyword, the position on the document, and the importance as an index, and the position on the document corresponding to the keyword are of importance. Since the related part presentation means and the index use device that present the keywords in descending order are configured, the positions of the keywords on the document can be presented in descending order of importance, which allows the user to see different places in the document many times. It is possible to quickly access the desired information.

【0081】

The invention according to claim 7 specifies a keyword designation means for designating a keyword of interest from a document having a keyword, a position on a document, and a importance as an index, and a lower limit of importance for presenting a related part. Since the index utilization device is configured with the minimum importance designation means and the related part presentation means that presents only those having the importance equal to or higher than the minimum importance at the position on the document corresponding to the keyword, the corresponding document of the keyword is configured. It is possible to present only those having an importance equal to or higher than the minimum importance specified by the position of, and thus it is possible to eliminate the need to access almost irrelevant parts.

【0082】

The invention according to claim 8 is a keyword designation means for designating a keyword of interest from a document having a keyword, a position on a document, and an importance as an index, and the number of presentation-related parts for one keyword. Since the index utilization device is configured with the means for specifying the maximum number of related parts that specifies the upper limit, the means for presenting the related parts that present the positions on the document corresponding to the keyword from the ones with the highest importance by the number less than the maximum number of related parts. , The position on the document corresponding to the keyword can be presented as many as the maximum number of related parts or less, which enables efficient display of related parts on a narrow screen at the same time or even when accessed through a slow communication path. It can be carried out.

[Simple explanation of drawings]

[Figure 1]

It is a block diagram which shows the structure of the index making apparatus which is one Example of the invention of claims 1 to 4.

[Figure 2]

It is a flowchart which shows the operation process of FIG.

[Fig. 3]

It is a block diagram which shows the structure of the index making apparatus which is one Example of the invention of claim 5.

[Fig. 4]

It is a flowchart which shows the operation process of FIG.

[Fig. 5]

It is a block diagram which shows the structure of the index utilization apparatus which is one Example of the invention of claims 6-8.

[Fig. 6]

It is a flowchart which shows the operation process of FIG.

[Explanation of symbols]

1 Keyword extraction method 2 Extraction rule storage means 3 Keyword selection method 4 Index element storage means 5 Keyword storage means 6 Position conversion means 7 Index generation means 8 Keyword extraction method 9 Keyword storage means 10 Position conversion means 11 Index generation means 12 Keyword specification means 13 Related points presentation means 14 Minimum importance designation means 15 Means for specifying the maximum number of related locations